- Email: [email protected]
The prognostic value of heart rate recovery in patients with coronary artery disease: A systematic review and meta-analysis
Accepted Manuscript The prognostic value of heart rate recovery in patients with coronary artery disease A systematic review and meta-analysis
Sangeeta Lachman, Michel S. Terbraak, Jacqueline Limpens, Harald Jorstad, Cees Lucas, Wilma Scholte op Reimer, S. Matthijs Boekholdt, Gerben ter Riet, Ron J.G. Peters PII: DOI: Reference:
S0002-8703(18)30051-6 https://doi.org/10.1016/j.ahj.2018.02.008 YMHJ 5630
To appear in: Received date: Accepted date:
6 June 2017 7 February 2018
Please cite this article as: Sangeeta Lachman, Michel S. Terbraak, Jacqueline Limpens, Harald Jorstad, Cees Lucas, Wilma Scholte op Reimer, S. Matthijs Boekholdt, Gerben ter Riet, Ron J.G. Peters , The prognostic value of heart rate recovery in patients with coronary artery disease A systematic review and meta-analysis. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Ymhj(2018), https://doi.org/10.1016/j.ahj.2018.02.008
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT The prognostic value of heart rate recovery in patients with coronary artery disease A systematic review and meta-analysis
Sangeeta Lachman MDa, Michel S. Terbraak MScb, Jacqueline Limpens PhDd, Harald Jorstad MD PhDa,
MDb PhD1, Ron J.G. Peters MD PhD1
RI
Affiliations:
PT
Cees Lucas PhDe, Wilma Scholte op Reimer RN, PhDb, S. Matthijs Boekholdta, Gerben ter Riet MD PhD
SC
a. Department of Cardiology, Academic Medical Center, Amsterdam, the Netherlands b. Achieve Amsterdam University of Applied Sciences, Amsterdam, the Netherlands
NU
c. Department of General Practice, Academic Medical Center, Amsterdam, the Netherlands
MA
d. Medical Library, Academic Medical Center, Amsterdam, the Netherlands e. Department of Biostatistics, Academic Medical Center, Amsterdam, the Netherlands
PT E
D
Key words: Coronary artery disease - heart rate recovery – mortality
Word counts manuscript: 4223 (references are included)
CE
Tables: 6, Figures: 2
S. Lachman
AC
Corresponding author:
Academic Medical Center Department of Cardiology Meibergdreef 9 1105 AZ Amsterdam, The Netherlands Email: [email protected] 1
ACCEPTED MANUSCRIPT Abstract Background Routine outpatient care of patients with coronary artery disease (CAD) lacks a simple measure of physical fitness and risk of mortality. Heart rate recovery (HRR) is noninvasive and easily obtainable in outpatient settings. Prior studies have suggested that delayed post-exercise HRR in the first
PT
minutes is associated with mortality in several types of populations. However, a comprehensive
RI
overview of the prognostic value of delayed HRR for time to mortality specifically in CAD patients is
SC
not available. The purpose of the current meta-analysis is to evaluate the prognostic value of delayed HRR in CAD patients.
NU
Methods
We conducted a systematic search in OVID MEDLINE and OVID EMBASE to identify studies reporting
MA
on HRR and risk of incident cardiovascular events or mortality in CAD patients. Hazard ratios for delayed vs non-delayed HRR were pooled using random effects meta-analysis.
D
Results
PT E
Four studies were included, comprising 2428 CAD patients. The study quality of the included studies was rated moderate (n=2) to high (n=2). Delayed HRR was defined by ≤12 to ≤21 bpm in the recovery
CE
period. During follow-up (range 2.0-9.8 years), 151 patients died (6.2% (range 2.5%-19.5%)). Only data on mortality could be pooled. Heterogeneity was limited (I2=32%; p=0.23), pooled unadjusted
Conclusion
AC
hazard ratio for mortality, based on 3 studies, was 5.8 (95%CI 3.2-10.4).
In CAD patients, delayed HRR is significantly associated with all-cause mortality. As exercise testing is performed routinely in CAD patients, HRR can be considered in monitoring exercise, still further research must investigate the addition of HRR in current risk scores. Wordcount: 251
2
ACCEPTED MANUSCRIPT Introduction Patients with coronary artery disease (CAD) are at high risk of recurrent events, with recurrence rates of up to 75% within three years (1, 2). Physical inactivity is a major risk factor for recurrence in CAD patients and is highly prevalent (range 35 to 60%) (3-5). A broad array of exercise tests is available in clinical practice, with cardiopulmonary exercise
PT
testing considered to be the gold standard for the assessment of physical fitness. However, a quick
RI
and objective measure for physical fitness that can be conducted during routine outpatient care is
SC
lacking.
Engagement in exercise improves autonomic nervous system function (6-10). Prior studies
NU
reported that impaired autonomic nervous system after peak exercise is an indicator of adverse cardiovascular outcomes in apparently healthy individuals (11-17). Autonomic function may be
MA
evaluated by post-exercise heart rate recovery (HRR) after exercise. HRR in the first minute of recovery after cessation of exercise reflects parasympathetic reactivation along with reduced
D
sympathetic activation during second to fifth minute of recovery. HRR is a noninvasive and less time
PT E
consuming measure compared to other exercise parameters that require advanced lab equipment such as oxygen peak by gas exchange measurement; furthermore HRR can be quickly obtained from
CE
routine exercise testing compared to other heart rate responses, such as heart rate variability measured with Holter electrocardiography (18).
AC
However, little is known about the prognostic value and optimal cut-off point of delayed HRR specifically in CAD patients and evaluation of HRR is not standard practice in outpatient clinics. The purpose of this systematic review and meta-analysis was to summarize the evidence about the prognostic value of delayed HRR in CAD patients, and to study reasons for heterogeneity, if any.
3
ACCEPTED MANUSCRIPT Methods The current systematic review was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO 2017: CRD42017054363 available on https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=54363) Source(s) of funding used to support the research and creation of the paper
PT
No extramural funding was used to support this work. The authors are solely responsible for the
RI
design and conduct of this study, all study analyses, the drafting and editing of the paper and its final
SC
contents.
Literature Search Strategy
NU
We followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (19). A medical librarian (JL) performed a systematic search in OVID MEDLINE and OVID
MA
EMBASE from inception to December 14th, 2016. We used controlled terms, such as MeSH (MEDLINE), and text words. We comprehensively searched for HRR and exercise combined with [i]
PT E
D
different types of CAD [CAD-I], [ii] CAD-interventions [CAD-II], [iii] terms that indicate CAD referral or rehabilitation [CAD-III] OR [iv], a narrow search for CVD (see appendix for the complete MEDLINE search strategy). No language or other restrictions were applied. The search included an iterative
CE
process to refine the search strategy by adding search terms if new relevant citations were identified by checking reference lists and cited articles using ISI Web of Science. ENDNOTE ® software (version
AC
12.0) was used to manage and deduplicate references. Inclusion and exclusion criteria We included prospective studies that reported on the association of HRR with incident cardiovascular events, cardiovascular mortality or all-cause mortality among CAD patients. Incident cardiovascular events were defined as ST-elevation myocardial infarction, non ST-elevation myocardial infarction, unstable angina pectoris, stable angina pectoris, coronary artery bypass surgery and heart failure. 4
ACCEPTED MANUSCRIPT We defined exclusion criteria as follows: a history of heart failure, a history of pacemaker implantation, and HRR obtained from exercise testing with the administration of drugs, such as adenosine, that could affect HRR. Selection of Articles Two investigators (SL and MT) independently screened all titles and abstracts, Discrepancies were
PT
resolved by consensus after review by a third investigator (GtR). Potentially eligible studies were
RI
retrieved and reviewed in full text. When multiple articles for a cohort population were present, we
SC
included the article with the largest study sample or with the most elaborate description of data. We contacted original study authors in order to obtain missing data on outcomes and statistical
NU
analyses.
MA
We used the Quality in Prognostic Studies (QUIPS) tool, provided by the Cochrane collaboration (20), for quality assessment of the included studies. The QUIPS-tool is based on six domains namely 1) study participants, 2) study attrition, 3) prognostic factor measurement, 4) outcome measurement,
PT E
D
5) study confounding, and 6) statistical analysis and reporting. We used five out of six domains of the QUIPS-tool; the domain ‘’study confounding’’ was omitted, since confounding is not relevant in the assessment of the prognostic value of HRR (21). Each domain was rated as high, moderate or low risk
CE
of bias. We judged overall risk of bias in each study as follows: low risk of bias was defined as ≥ 4 criteria graded as low risk of bias and none of the criteria graded as high risk of bias, moderate risk of
AC
bias was defined as ≥ 2 criteria rated as low risk of bias and the remaining criteria rated as moderate risk of bias, and high risk of bias was defined as ≥ 1 criteria rated as high risk of bias. Data Extraction The following information was retrieved from the included studies: country of study population, year of publication; study design; number of participants at baseline; total number of events, based on nonfatal and fatal cardiovascular events and total mortality; duration of follow-up; study outcome 5
ACCEPTED MANUSCRIPT measured in number of events and/or mortality, defined as either all-cause mortality or cardiovascular death; age in mean or median and standard deviation (SD) or interquartile range (IQR); proportion of men at baseline; proportion of patients on beta-blockers, calcium channel blockers and/or digitalis at baseline; resting heart rate in beats per minute (bpm); the proportion of patients with hypertension and/or diabetes mellitus at baseline; body mass index (BMI) in kg/m2; the
PT
proportion of smokers at baseline; mean left ventricular ejection fraction (LVEF) (%); the proportion of patients with atrial fibrillation; measurement characteristics of HRR such as type of exercise
RI
(treadmill or bicycle); level of exercise intensity (maximal or submaximal); exercise position (seated /
SC
standing); passive or active recovery; recovery period of heart rate in minutes; cutoff of delayed HRR;
NU
statistical model; association measure, such as hazard ratio/risk ratio/odds ratio and confidence intervals (CI); and number of covariates which were entered into the multivariable prediction model.
MA
If studies dichotomized HRR into delayed vs not delayed and also used continuous HRR in statistical analysis, the most elaborated results were chosen in the presentation of data extraction. Two
D
investigators (SL and MT) independently conducted the data extraction using a protocolled template.
PT E
Data extraction of one investigator (SL) was checked by the second investigator (MT). Disagreements were adjudicated by a third reader (GtR).
CE
Methods of analysis
For each regression model we extracted hazards ratio/risk ratio/odds ratio to study the associations
AC
between physical fitness / autonomic function and the outcome. Heterogeneity was assessed visually and by testing for heterogeneity (based on the I2 statistic). P-value of <0.05 was considered as significant for heterogeneity. We used a random effects meta-analysis to estimate the overall association between HRR and outcome, which was based on univariable hazard ratio, if available. Like the original authors we assumed that hazard ratios were constant over time. Statistical analyses were performed in Stata version 13.1 (Stata Corporation, College station, TX, USA).
6
ACCEPTED MANUSCRIPT Results After reviewing 2078 titles and abstracts, 66 studies were assessed full text for eligibility. In total, four studies met the inclusion criteria (figure 1) (7, 18, 22, 23). No additional studies were identified by reference checking. Table 1 shows the study characteristics. A total number of 2428 CAD patients were included in these four studies, with 151 outcome events during follow-up (range 2.0-9.8 years).
PT
Events were reported as all-cause mortality (18, 22), and cardiac mortality (7), with one study reporting a composite outcome, including cardiovascular death and hospitalization for congestive
RI
heart failure (23).
SC
Study population
Table 2 shows baseline characteristics of the study populations. The mean age ranged between
NU
61±11 (SD) to 67±9 (SD) years, 73% (1772/2428) were male. The four included studies did not report
MA
on the impact of beta-blockers on HRR separately. In total, 91% (2218/2428) of the study populations were treated with beta-blockers. The proportion of diabetes mellitus was 36% (877/2428). The proportion of hypertension, reported in 3 studies, was 51% (455/897). BMI, reported in 2 studies,
PT E
D
ranged from 28 to 29 kg/m2. The use of calcium channel blockers, digitalis and the prevalence of atrial fibrillation in the studies was reported inconsistently and incompletely and is not reported
CE
here. Study characteristics
AC
Characteristics of the assessment of HRR are presented in table 3. HRR was measured 1 minute postexercise in all 4 studies and the cutoff value for delayed HRR ranged from 12 to 21 beats in the first minute. The Tables 4a and 4b show the statistical methods used in the studies. The adjusted hazard ratio for delayed HRR vs not delayed HRR was reported in all four studies, whereas the unadjusted hazard ratio was reported in three out of four studies. The nature and number of covariates in the various models varied greatly. Detailed information on covariates which were entered into the statistical model is presented in table 4b. 7
ACCEPTED MANUSCRIPT Primary outcome The unadjusted hazard ratio and the pooled unadjusted hazard ratio for the risk of all-cause mortality and composite outcome (cardiovascular death and hospitalization for congestive heart failure), was 5.78 (95%CI 3.20-10.43) (figure 2). I-squared was 32% (p= 0.23), indicating moderate, but statistically non-significant heterogeneity. Table 5 shows the results of the risk of bias assessment of the
PT
included studies; two studies were rated as high quality (22, 23) and two studies were rated as
RI
moderate quality (7, 18).
SC
Discussion
Our meta-analysis, using data from 2484 CAD patients (151 deaths) shows that delayed HRR after
NU
exercise is associated with an increase in all-cause mortality in this patient population. The
MA
incorporation of HRR in routine exercise testing may improve the monitoring of exercise tolerance and risk stratification for mortality in CAD patients.
In general, modifiable cardiovascular risk factors such as overweight, smoking and
PT E
D
cholesterol status are objectively measured using BMI, cotinine testing, and plasma lipid levels. As HRR is more easily measured than other autonomic indices and exercise parameters (18), HRR may be useful as a quick and objective prognostic tool. Particularly, HRR from submaximal exercise test
CE
could potentially be used in monitoring physical fitness. Of the studies included in our meta-analysis, Nissinen et al. reported that HRR is a good
AC
predictor of death compared to traditional autonomic markers such as heart rate variability and baroreflex sensitivity (18). However, Kivinimie et al. found a similar association between HRR and cardiovascular death and hospitalization and maximal chronotropic response index, with an adjusted hazard ratio of 6.0 (2.5-14.5) and 4.9 (2.1-11.3), respectively (23). In the current meta-analysis, the pooled hazard ratio for the risk of all-cause, cardiovascular mortality and composite adverse cardiovascular events was 5.8 (3.2-19.4), which supports stratification of CAD patients according to the risk of mortality. Furthermore, our findings were consistent with a large meta-analysis in 8
ACCEPTED MANUSCRIPT apparently healthy individuals, where delayed HRR was associated with an increased risk of cardiovascular events and all-cause mortality(17). Qiu et al. support the use of HRR for risk assessment in clinical practice. The majority of CAD patients are treated with beta-blockers, potentially influencing HRR. Desai et al. reported that patients treated with beta-blockers have slower HRR compared to patients
PT
without beta-blockers, but they did not report the impact on hazard ratios (24). Considering the pooled hazard ratio of 5.8 in our study, the prognostic value of the HRR for CAD patients seems to
RI
stay intact, notwithstanding the fact that 91% of all patients in our analysis were on beta-blocker
SC
treatment.
NU
The underlying mechanism of impaired HRR
The underlying mechanism of impaired autonomic regulation after peak exercise is not
MA
completely understood. Delayed reactivation of parasympathetic nervous system after cessation of exercise may be induced by a highly activated sympathetic nervous system. Biological links with
D
mortality include impaired vagal reactivation, pro-inflammatory activation, decreased endothelial
PT E
function and arterial stiffness (25-27). Additionally, the prognostic value of HRR has also been observed in other diseases with risk of impaired autonomic responses, such as heart failure, diabetes
CE
mellitus and cancer (27-31). These diseases have in common that stress and inflammatory factors may impair the autonomic function. These stress and inflammatory factors may have a potential role
AC
in the underlying mechanism of impaired HRR in our study population as well. Recovery period and exercise intensity levels HRR has been shown to be prognostic at 1, 2, and 5 minutes after exercise in different populations referred for exercise testing (10, 11, 29, 31, 32). Cole et al. demonstrated that the cutoff value of 12 beats in the first minute of recovery provides maximal discrimination as a prognostic marker for mortality (11), which suggests that parasympathetic reactivity may be more predictive of mortality than sympathetic withdrawal during the later stage of recovery. According to our findings, 9
ACCEPTED MANUSCRIPT the cutoff value of HRR may range from 12 to 21 beats in the first minute of recovery. The cutoff value of 21 beats with a high hazard ratio by Kiviniemi et al. compared to the other included studies in the meta-analysis, may be explained by the difference in outcome measurement, particularly, the use of a composite outcome of mortality and hospitalization of which the latter was not included in the other studies.
PT
In our analysis HRR was derived from symptom-limited cardiopulmonary exercise testing and alternatives such as submaximal exercise test should be considered. In a cohort of 5,234 apparently
RI
healthy individuals, the Lipid Research Clinics Prevalence Study demonstrated that delayed HRR
SC
derived from submaximal exercise testing is significantly associated with mortality (33). Therefore,
NU
submaximal exercise tests might be also useful in monitoring risks and physical fitness at physician check-ups or during cardiac rehabilitation. However, submaximal tests in CAD patients still need to
MA
be validated. Strength and limitations
D
The main strength of this systematic review and meta-analysis was the systematic approach,
PT E
literature search, systematic grading of studies and data extraction. We systematically assessed the statistical analyses in all studies and the prognostic value of HRR was broadly estimated. Limitations
CE
should be noted as well. First, heterogeneity did not allow pooling of individual-level patient data. Second, in one of the studies, the unadjusted hazard ratio was not available even after contacting the
AC
authors (22). Third, there was heterogeneity in co-variates, study outcome, exercise and recovery protocols such as the application of either treadmill testing or the cycle ergometer to measure HRR in the included studies. Since there was heterogeneity in co-variates in the prediction models, we used the unadjusted hazard ratio in the meta-analysis. Furthermore, the heterogeneity of study outcome was due to a composite outcome in one study, which was based on cardiovascular death and hospitalization for congestive heart failure (23). However, all four studies studied mortality as an outcome, and in 2 out of 4 all-cause mortality was studied. And, although there was heterogeneity in 10
ACCEPTED MANUSCRIPT exercise protocols which may lead to differences in maximal workloads, this may not be clinically relevant as the value of HRR as a prognostic marker has been reported to be independent of workload (11, 33). Fourth, chronotropic incompetence and low exercise tolerance, which have also been shown to be associated with the risk of mortality (34, 35), may bias the association between HRR and the risk of mortality. However, according to Kiviniemie et al. HRR independently predicts all-
PT
cause mortality even in the presence of the aforementioned factors (23).
RI
Future research should focus on the added value of HRR in clinical prediction models in large
SC
CAD populations compared to exercise parameters from the Duke treadmill score. Prediction models should have good discrimination of event and no event and calibration (predicted probabilities
NU
should match those observed). Furthermore, since HRR is often derived from symptom-limited exercise testing in CAD patients, the value of HRR derived from submaximal exercise testing as a
MA
metric of physical fitness and an independent prognostic marker in CAD patients needs to be established.
PT E
D
Conclusion
In CAD patients delayed HRR is significantly associated with overall mortality, and provides clinically additional information in this patient population. As HRR is a noninvasive measure of autonomic
CE
function, HRR can be considered in the monitoring of exercise tolerance. Still, future studies are needed to clarify its applicability in non-laboratory settings, in submaximal exercise tests and in
AC
current risk scores.
Acknowledgements
No funders had a role in the study design, data collection, analysis, decision to publish or preparation of the manuscript. Funding sources None 11
ACCEPTED MANUSCRIPT Author contributions SL, MST and RP contributed to the conception and design of the work. SL, MST, JL and GT contributed to the acquisition, analysis and interpretation of data for the work. SL, and MST drafted the manuscript. JL, HT, WSR, CL, SMB, GT and RP critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy.
PT
Disclosures
RI
SMB reported personal fees from Pfizer, outside the submitted work; RJG reported personal fees
SC
from Sanofi Aventis, Boehringer Ingelheim, Amgen and AstraZeneca, outside the submitted work.
AC
CE
PT E
D
MA
NU
The other authors declare that they have no known conflicts of interest.
12
ACCEPTED MANUSCRIPT References:
AC
CE
PT E
D
MA
NU
SC
RI
PT
1. Sixty-Sixth World Health Assembly. Follow-up to the Political Declaration of the High-level Meeting of the General Assembly on the Prevention and Control of Non-communicable Diseases. WHA66.10, 27 May 2013. http://apps.who.int/gb/ebwha/pdf_files/WHA66/A66_R10-en.pdf. 2. Bansilal S, Castellano JM, Fuster V. Global burden of CVD: focus on secondary prevention of cardiovascular disease. Int J Cardiol. 2015;201 Suppl 1:S1-7. 3. Kotseva K, Wood D, De Bacquer D, De Backer G, Ryden L, Jennings C, et al. EUROASPIRE IV: A European Society of Cardiology survey on the lifestyle, risk factor and therapeutic management of coronary patients from 24 European countries. Eur J Prev Cardiol. 2015. 4. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT, et al. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838):219-29. 5. Teo K, Lear S, Islam S, Mony P, Dehghan M, Li W, et al. Prevalence of a healthy lifestyle among individuals with cardiovascular disease in high-, middle- and low-income countries: The Prospective Urban Rural Epidemiology (PURE) study. JAMA. 2013;309(15):1613-21. 6. Anderson L, Oldridge N, Thompson DR, Zwisler AD, Rees K, Martin N, et al. Exercise-Based Cardiac Rehabilitation for Coronary Heart Disease: Cochrane Systematic Review and Meta-Analysis. J Am Coll Cardiol. 2016;67(1):1-12. 7. Hai JJ, Siu CW, Ho HH, Li SW, Lee S, Tse HF. Relationship between changes in heart rate recovery after cardiac rehabilitation on cardiovascular mortality in patients with myocardial infarction. Heart Rhythm. 2010;7(7):929-36. 8. Lucini D, Milani RV, Costantino G, Lavie CJ, Porta A, Pagani M. Effects of cardiac rehabilitation and exercise training on autonomic regulation in patients with coronary artery disease. Am Heart J. 2002;143(6):977-83. 9. Malfatto G, Facchini M, Sala L, Branzi G, Bragato R, Leonetti G. Relationship between baseline sympatho-vagal balance and the autonomic response to cardiac rehabilitation after a first uncomplicated myocardial infarction. Ital Heart J. 2000;1(3):226-32. 10. Ribisl PM, Gaussoin SA, Lang W, Bahnson J, Connelly SA, Horton ES, et al. Lifestyle intervention improves heart rate recovery from exercise in adults with type 2 diabetes: results from the Look AHEAD study. J Obes. 2012;2012:309196. 11. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate recovery immediately after exercise as a predictor of mortality. NEnglJMed. 1999;341(18):1351-7. 12. Girotra S, Keelan M, Weinstein AR, Mittleman MA, Mukamal KJ. Relation of heart rate response to exercise with prognosis and atherosclerotic progression after coronary artery bypass grafting. Am J Cardiol. 2009;103(10):1386-90. 13. Ho JS, Fitzgerald SJ, Barlow CE, Cannaday JJ, Kohl HW, 3rd, Haskell WL, et al. Risk of mortality increases with increasing number of abnormal non-ST parameters recorded during exercise testing. Eur J Cardiovasc Prev Rehabil. 2010;17(4):462-8. 14. Lauer MS, Okin PM, Larson MG, Evans JC, Levy D. Impaired heart rate response to graded exercise. Prognostic implications of chronotropic incompetence in the Framingham Heart Study. Circulation. 1996;93(8):1520-6. 15. Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and sudden cardiac death. Experimental basis and clinical observations for post-myocardial infarction risk stratification. Circulation. 1992;85(1 Suppl):I77-91. 16. Shetler K, Marcus R, Froelicher VF, Vora S, Kalisetti D, Prakash M, et al. Heart rate recovery: validation and methodologic issues. J Am Coll Cardiol. 2001;38(7):1980-7. 17. Qiu S, Cai X, Sun Z, Li L, Zuegel M, Steinacker JM, et al. Heart Rate Recovery and Risk of Cardiovascular Events and All-Cause Mortality: A Meta-Analysis of Prospective Cohort Studies. Journal of the American Heart Association. 2017;6(5). 13
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
18. Nissinen SI, Makikallio TH, Seppanen T, Tapanainen JM, Salo M, Tulppo MP, et al. Heart rate recovery after exercise as a predictor of mortality among survivors of acute myocardial infarction. Am J Cardiol. 2003;91(6):711-4. 19. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6(7):e1000100. 20. Hayden JA, van der Windt DA, Cartwright JL, Cote P, Bombardier C. Assessing bias in studies of prognostic factors. Ann Intern Med. 2013;158(4):280-6. 21. Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594. 22. Aijaz B, Squires RW, Thomas RJ, Johnson BD, Allison TG. Predictive value of heart rate recovery and peak oxygen consumption for long-term mortality in patients with coronary heart disease. Am J Cardiol. 2009;103(12):1641-6. 23. Kiviniemi AM, Lepojarvi S, Kentta TV, Junttila MJ, Perkiomaki JS, Piira OP, et al. Exercise capacity and heart rate responses to exercise as predictors of short-term outcome among patients with stable coronary artery disease. Am J Cardiol. 2015;116(10):1495-501. 24. Desai MY, De la Pena-Almaguer E, Mannting F. Abnormal heart rate recovery after exercise as a reflection of an abnormal chronotropic response. Am J Cardiol. 2001;87(10):1164-9. 25. Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405(6785):458-62. 26. Jae SY, Heffernan KS, Park SH, Jung SH, Yoon ES, Kim EJ, et al. Does an acute inflammatory response temporarily attenuate parasympathetic reactivation? Clin Auton Res. 2010;20(4):229-33. 27. Youn JC, Lee HS, Choi SW, Han SW, Ryu KH, Shin EC, et al. Post-Exercise Heart Rate Recovery Independently Predicts Clinical Outcome in Patients with Acute Decompensated Heart Failure. PLoS One. 2016;11(5):e0154534. 28. Cheng YJ, Lauer MS, Earnest CP, Church TS, Kampert JB, Gibbons LW, et al. Heart rate recovery following maximal exercise testing as a predictor of cardiovascular disease and all-cause mortality in men with diabetes. Diabetes Care. 2003;26(7):2052-7. 29. Giallauria F, Maresca L, Vitelli A, Santucci de Magistris M, Chiodini P, Mattiello A, et al. Exercise training improves heart rate recovery in women with breast cancer. Springerplus. 2015;4:388. 30. Ha D, Choi H, Zell K, Raymond DP, Stephans K, Wang XF, et al. Association of impaired heart rate recovery with cardiopulmonary complications after lung cancer resection surgery. J Thorac Cardiovasc Surg. 2015;149(4):1168-73 e3. 31. Minai OA, Nguyen Q, Mummadi S, Walker E, McCarthy K, Dweik RA. Heart rate recovery is an important predictor of outcomes in patients with connective tissue disease-associated pulmonary hypertension. Pulm Circ. 2015;5(3):565-76. 32. Goda A, Koike A, Iwamoto MH, Nagayama O, Yamaguchi K, Tajima A, et al. Prognostic value of heart rate profiles during cardiopulmonary exercise testing in patients with cardiac disease. Int Heart J. 2009;50(1):59-71. 33. Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after submaximal exercise testing as a predictor of mortality in a cardiovascularly healthy cohort. AnnInternMed. 2000;132(7):552-5. 34. Jouven X, Empana JP, Schwartz PJ, Desnos M, Courbon D, Ducimetiere P. Heart-rate profile during exercise as a predictor of sudden death. N Engl J Med. 2005;352(19):1951-8. 35. Savonen KP, Kiviniemi V, Laukkanen JA, Lakka TA, Rauramaa TH, Salonen JT, et al. Chronotropic incompetence and mortality in middle-aged men with known or suspected coronary heart disease. Eur Heart J. 2008;29(15):1896-902. 14
ACCEPTED MANUSCRIPT Table 1. Characteristics of included studies Kiviniemi et al. 2015 Finland
Country
Hai et al., 2010 China Cardiac rehabilitation
Study Source
Aijaz et al., 2009 USA Cardiac rehabilitation
Nissinen et al., 2003 Finland
Tot al
NU
SC
RI
PT
ARTEMIS Not documented Participants, n 242 (%) 1531 386 282 229 8 Follow-up, years 2.0 6.5 ± 3.4 9.8 ± 2.9 2.6 ± 1.1 Primary Composite All-cause All -cause outcome outcome* Cardiac death* mortality mortality Total events*, n (%) 39 (2.5) 40 (10.4) 55 (19.5) 17 (7.4) 151 Average number of events per year 20 6 6 7 ARTEMIS=Innovation to Reduce Cardiovascular Complications of Diabetes at the Intersection * composite endpoint indicates cardiovascular death and hospitalization for congestive heart failure * cardiac death indicates death due to myocardial infarction, heart failure, or malignant arrhythmia * Total events indicate all-cause, cardiovascular mortality and composite adverse cardiovascular events
Male, n (%)
Resting heart rate (mean±SD)
Hypertension, n (%)
AC
Diabetes mellitus, n (%) Body mass index (mean±SD)
D
Hai et al., 2010 386 (100) 65.2 ± 11.5
Aijaz et al., 2009 282(100) 61.0 ± 11.0
Nissinen et al., 2003 229 (100) 61.5 ± 9.5
1054 (69)
297 (77)
234 (83)
187 (82)
1531 (100)
304 (79)
154 (55)
229 (100)
63.0 ± 12.0
72.0 ± 14.0
71.0 ± 13.0
N/A
CE
Beta-blockers, n (%)
PT E
CAD, n (%) Age (years, mean±SD)
Kiviniemi et al. 2015 1531 (100) 67.0 ± 8.7
MA
Table 2. Baseline characteristics of study participants
N/A
208 (54)
143 (5)
104 (45)
660 (43)
147 (38)
36 (13)
34 (15)
29.0 ± 4.7
N/A
27.6 ± 4.3
N/A
Total (%) 2428 (100) 1.772/2428 (73) 2.218/2428 (91)
455/897 (51) 877/2428 (36)
340/2199 134 (9) 148 (38) 58 (21) N/A (15) LVEF (mean±SD) 62.0 ± 11 45.4 ± 10.0 52.0 ± 13 46.0 ± 9.0 CAD=coronary artery disease, LVEF=left ventricular ejection fraction, N/A=not available and SD=standard deviation. Smoking, n (%)
Table 3. Measurement characteristics of heart rate recovery 15
ACCEPTED MANUSCRIPT
Study
Exercise
Measurement position
Recovery
Recovery (min)
HRR cutoff
Kiviniemi et al., 2015 BET Supine/standing Passive 1 21 Hai et al., 2010 TET Seated Passive 1 12 Aijaz et al., 2009 TET N/A Active 1 13 Nissinen et al., 2003 BET Seated Passive 1 12 All exercise tests were based on maximal level of intensity. BET=bicycle exercise test, TET=treadmill exercise test and HRR=heart rate recovery. The passive recovery means absence of cooling-down after cessation of exercise testing.
Statistical analysis
HRR cutoff
unadjusted HR (95% CI)
adjusted HR (95% CI)
SC
RI
Study outcome Composite U/M Cox ph 21 9.9 (4.4-22.5) 6.0 (2.5-14.5) outcome* U/M Cox ph 12 4.0 (2.1-7.5) 2.5 (1.1-5.6) Cardiac death* All-cause Aijaz et al., 2009 M Cox ph 13 N/A 2.2 (1.1-4.1) mortality Nissinen et al., All-cause 2003 U/M Cox ph 12 5.3 (1.5-18.5) 4.3 (1.2-16.0) mortality U/M Cox ph=univariable or multivariable Cox proportional hazards model, CI=confidence interval, HRR=heart rate recovery, HR=hazard ratio, * composite outcome indicates cardiovascular death and hospitalization for congestive heart failure * cardiac death indicates death due to myocardial infarction, intractable heart failure, or malignant arrhythmia.
AC
CE
PT E
D
MA
NU
Study Kiviniemi et al., 2015 Hai et al., 2010
PT
Table 4a. Statistical analysis
16
ACCEPTED MANUSCRIPT Table 4b. Statistical analysis Study Name Number of covariates in final model Covariates Kiviniemi et al., 2015 10 Age, sex, diabetes mellitus, smoking status, BMI, history of acute myocardial infarction, history of coronary intervention, Canadian Cardiovascular Society grading of angina pectoris, LVEF and resting heart rate. 14
Aijaz et al., 2009
3
Diabetes mellitus, phase 1 LVEF ≤30%, phase 2 LVEF≤30%, acute pulmonary edema, revascularizations before enrollment, use of angiotensinconverting enzyme inhibitor/angiotensin receptor blocker, statin, exercise training, exercise parameters including phase 1 resting heart rate≥65 bpm, phase 1 heart rate increment ≤45bpm, phase 1 exercise capacity, phase duke treadmill score, phase 2 heart rate increment≤45bpm, phase 2 HRR < 12 bpm and phase 2 exercise capacity ≤4 METs.
MA
NU
SC
RI
PT
Hai et al., 2010
Age, sex, low peak VO2.
PT E
D
3 Nissinen et al., 2003 Age, diabetes mellitus and NYHA class BMI=body mass index, LVEF=left ventricular ejection fraction, HRR=heart rate recovery, METs=metabolic equivalents, peak VO2=peak oxygen consumption and NYHA indicates New York Heart Association. Table 5. Quality Assessment based on QUIPS defined by Cochrane Kiviniemi et al., 2015 low moderate low low low
CE
Domains
AC
Study Participants Study Attrition Prognostic factor Outcome Statistical analysis and reporting Overall Score
Hai et al., 2010 Aijaz et al., 2009 moderate low moderate low low low low low moderate moderate
Nissinen et al., 2003 moderate moderate low Moderate low
High quality
Moderate High quality Moderate quality quality QUIPS=Quality in Prognostic Studies tool, high quality ≥ 4 criteria low risk of bias and no criteria of high risk of bias, moderate quality ≥ 2 criteria low risk of bias and the remaining criteria were rated as moderate risk of bias. Low quality ≥ 1 criteria high risk of bias.
17
Figure 1
Figure 2
Sangeeta Lachman, Michel S. Terbraak, Jacqueline Limpens, Harald Jorstad, Cees Lucas, Wilma Scholte op Reimer, S. Matthijs Boekholdt, Gerben ter Riet, Ron J.G. Peters PII: DOI: Reference:
S0002-8703(18)30051-6 https://doi.org/10.1016/j.ahj.2018.02.008 YMHJ 5630
To appear in: Received date: Accepted date:
6 June 2017 7 February 2018
Please cite this article as: Sangeeta Lachman, Michel S. Terbraak, Jacqueline Limpens, Harald Jorstad, Cees Lucas, Wilma Scholte op Reimer, S. Matthijs Boekholdt, Gerben ter Riet, Ron J.G. Peters , The prognostic value of heart rate recovery in patients with coronary artery disease A systematic review and meta-analysis. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Ymhj(2018), https://doi.org/10.1016/j.ahj.2018.02.008
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT The prognostic value of heart rate recovery in patients with coronary artery disease A systematic review and meta-analysis
Sangeeta Lachman MDa, Michel S. Terbraak MScb, Jacqueline Limpens PhDd, Harald Jorstad MD PhDa,
MDb PhD1, Ron J.G. Peters MD PhD1
RI
Affiliations:
PT
Cees Lucas PhDe, Wilma Scholte op Reimer RN, PhDb, S. Matthijs Boekholdta, Gerben ter Riet MD PhD
SC
a. Department of Cardiology, Academic Medical Center, Amsterdam, the Netherlands b. Achieve Amsterdam University of Applied Sciences, Amsterdam, the Netherlands
NU
c. Department of General Practice, Academic Medical Center, Amsterdam, the Netherlands
MA
d. Medical Library, Academic Medical Center, Amsterdam, the Netherlands e. Department of Biostatistics, Academic Medical Center, Amsterdam, the Netherlands
PT E
D
Key words: Coronary artery disease - heart rate recovery – mortality
Word counts manuscript: 4223 (references are included)
CE
Tables: 6, Figures: 2
S. Lachman
AC
Corresponding author:
Academic Medical Center Department of Cardiology Meibergdreef 9 1105 AZ Amsterdam, The Netherlands Email: [email protected] 1
ACCEPTED MANUSCRIPT Abstract Background Routine outpatient care of patients with coronary artery disease (CAD) lacks a simple measure of physical fitness and risk of mortality. Heart rate recovery (HRR) is noninvasive and easily obtainable in outpatient settings. Prior studies have suggested that delayed post-exercise HRR in the first
PT
minutes is associated with mortality in several types of populations. However, a comprehensive
RI
overview of the prognostic value of delayed HRR for time to mortality specifically in CAD patients is
SC
not available. The purpose of the current meta-analysis is to evaluate the prognostic value of delayed HRR in CAD patients.
NU
Methods
We conducted a systematic search in OVID MEDLINE and OVID EMBASE to identify studies reporting
MA
on HRR and risk of incident cardiovascular events or mortality in CAD patients. Hazard ratios for delayed vs non-delayed HRR were pooled using random effects meta-analysis.
D
Results
PT E
Four studies were included, comprising 2428 CAD patients. The study quality of the included studies was rated moderate (n=2) to high (n=2). Delayed HRR was defined by ≤12 to ≤21 bpm in the recovery
CE
period. During follow-up (range 2.0-9.8 years), 151 patients died (6.2% (range 2.5%-19.5%)). Only data on mortality could be pooled. Heterogeneity was limited (I2=32%; p=0.23), pooled unadjusted
Conclusion
AC
hazard ratio for mortality, based on 3 studies, was 5.8 (95%CI 3.2-10.4).
In CAD patients, delayed HRR is significantly associated with all-cause mortality. As exercise testing is performed routinely in CAD patients, HRR can be considered in monitoring exercise, still further research must investigate the addition of HRR in current risk scores. Wordcount: 251
2
ACCEPTED MANUSCRIPT Introduction Patients with coronary artery disease (CAD) are at high risk of recurrent events, with recurrence rates of up to 75% within three years (1, 2). Physical inactivity is a major risk factor for recurrence in CAD patients and is highly prevalent (range 35 to 60%) (3-5). A broad array of exercise tests is available in clinical practice, with cardiopulmonary exercise
PT
testing considered to be the gold standard for the assessment of physical fitness. However, a quick
RI
and objective measure for physical fitness that can be conducted during routine outpatient care is
SC
lacking.
Engagement in exercise improves autonomic nervous system function (6-10). Prior studies
NU
reported that impaired autonomic nervous system after peak exercise is an indicator of adverse cardiovascular outcomes in apparently healthy individuals (11-17). Autonomic function may be
MA
evaluated by post-exercise heart rate recovery (HRR) after exercise. HRR in the first minute of recovery after cessation of exercise reflects parasympathetic reactivation along with reduced
D
sympathetic activation during second to fifth minute of recovery. HRR is a noninvasive and less time
PT E
consuming measure compared to other exercise parameters that require advanced lab equipment such as oxygen peak by gas exchange measurement; furthermore HRR can be quickly obtained from
CE
routine exercise testing compared to other heart rate responses, such as heart rate variability measured with Holter electrocardiography (18).
AC
However, little is known about the prognostic value and optimal cut-off point of delayed HRR specifically in CAD patients and evaluation of HRR is not standard practice in outpatient clinics. The purpose of this systematic review and meta-analysis was to summarize the evidence about the prognostic value of delayed HRR in CAD patients, and to study reasons for heterogeneity, if any.
3
ACCEPTED MANUSCRIPT Methods The current systematic review was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO 2017: CRD42017054363 available on https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=54363) Source(s) of funding used to support the research and creation of the paper
PT
No extramural funding was used to support this work. The authors are solely responsible for the
RI
design and conduct of this study, all study analyses, the drafting and editing of the paper and its final
SC
contents.
Literature Search Strategy
NU
We followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (19). A medical librarian (JL) performed a systematic search in OVID MEDLINE and OVID
MA
EMBASE from inception to December 14th, 2016. We used controlled terms, such as MeSH (MEDLINE), and text words. We comprehensively searched for HRR and exercise combined with [i]
PT E
D
different types of CAD [CAD-I], [ii] CAD-interventions [CAD-II], [iii] terms that indicate CAD referral or rehabilitation [CAD-III] OR [iv], a narrow search for CVD (see appendix for the complete MEDLINE search strategy). No language or other restrictions were applied. The search included an iterative
CE
process to refine the search strategy by adding search terms if new relevant citations were identified by checking reference lists and cited articles using ISI Web of Science. ENDNOTE ® software (version
AC
12.0) was used to manage and deduplicate references. Inclusion and exclusion criteria We included prospective studies that reported on the association of HRR with incident cardiovascular events, cardiovascular mortality or all-cause mortality among CAD patients. Incident cardiovascular events were defined as ST-elevation myocardial infarction, non ST-elevation myocardial infarction, unstable angina pectoris, stable angina pectoris, coronary artery bypass surgery and heart failure. 4
ACCEPTED MANUSCRIPT We defined exclusion criteria as follows: a history of heart failure, a history of pacemaker implantation, and HRR obtained from exercise testing with the administration of drugs, such as adenosine, that could affect HRR. Selection of Articles Two investigators (SL and MT) independently screened all titles and abstracts, Discrepancies were
PT
resolved by consensus after review by a third investigator (GtR). Potentially eligible studies were
RI
retrieved and reviewed in full text. When multiple articles for a cohort population were present, we
SC
included the article with the largest study sample or with the most elaborate description of data. We contacted original study authors in order to obtain missing data on outcomes and statistical
NU
analyses.
MA
We used the Quality in Prognostic Studies (QUIPS) tool, provided by the Cochrane collaboration (20), for quality assessment of the included studies. The QUIPS-tool is based on six domains namely 1) study participants, 2) study attrition, 3) prognostic factor measurement, 4) outcome measurement,
PT E
D
5) study confounding, and 6) statistical analysis and reporting. We used five out of six domains of the QUIPS-tool; the domain ‘’study confounding’’ was omitted, since confounding is not relevant in the assessment of the prognostic value of HRR (21). Each domain was rated as high, moderate or low risk
CE
of bias. We judged overall risk of bias in each study as follows: low risk of bias was defined as ≥ 4 criteria graded as low risk of bias and none of the criteria graded as high risk of bias, moderate risk of
AC
bias was defined as ≥ 2 criteria rated as low risk of bias and the remaining criteria rated as moderate risk of bias, and high risk of bias was defined as ≥ 1 criteria rated as high risk of bias. Data Extraction The following information was retrieved from the included studies: country of study population, year of publication; study design; number of participants at baseline; total number of events, based on nonfatal and fatal cardiovascular events and total mortality; duration of follow-up; study outcome 5
ACCEPTED MANUSCRIPT measured in number of events and/or mortality, defined as either all-cause mortality or cardiovascular death; age in mean or median and standard deviation (SD) or interquartile range (IQR); proportion of men at baseline; proportion of patients on beta-blockers, calcium channel blockers and/or digitalis at baseline; resting heart rate in beats per minute (bpm); the proportion of patients with hypertension and/or diabetes mellitus at baseline; body mass index (BMI) in kg/m2; the
PT
proportion of smokers at baseline; mean left ventricular ejection fraction (LVEF) (%); the proportion of patients with atrial fibrillation; measurement characteristics of HRR such as type of exercise
RI
(treadmill or bicycle); level of exercise intensity (maximal or submaximal); exercise position (seated /
SC
standing); passive or active recovery; recovery period of heart rate in minutes; cutoff of delayed HRR;
NU
statistical model; association measure, such as hazard ratio/risk ratio/odds ratio and confidence intervals (CI); and number of covariates which were entered into the multivariable prediction model.
MA
If studies dichotomized HRR into delayed vs not delayed and also used continuous HRR in statistical analysis, the most elaborated results were chosen in the presentation of data extraction. Two
D
investigators (SL and MT) independently conducted the data extraction using a protocolled template.
PT E
Data extraction of one investigator (SL) was checked by the second investigator (MT). Disagreements were adjudicated by a third reader (GtR).
CE
Methods of analysis
For each regression model we extracted hazards ratio/risk ratio/odds ratio to study the associations
AC
between physical fitness / autonomic function and the outcome. Heterogeneity was assessed visually and by testing for heterogeneity (based on the I2 statistic). P-value of <0.05 was considered as significant for heterogeneity. We used a random effects meta-analysis to estimate the overall association between HRR and outcome, which was based on univariable hazard ratio, if available. Like the original authors we assumed that hazard ratios were constant over time. Statistical analyses were performed in Stata version 13.1 (Stata Corporation, College station, TX, USA).
6
ACCEPTED MANUSCRIPT Results After reviewing 2078 titles and abstracts, 66 studies were assessed full text for eligibility. In total, four studies met the inclusion criteria (figure 1) (7, 18, 22, 23). No additional studies were identified by reference checking. Table 1 shows the study characteristics. A total number of 2428 CAD patients were included in these four studies, with 151 outcome events during follow-up (range 2.0-9.8 years).
PT
Events were reported as all-cause mortality (18, 22), and cardiac mortality (7), with one study reporting a composite outcome, including cardiovascular death and hospitalization for congestive
RI
heart failure (23).
SC
Study population
Table 2 shows baseline characteristics of the study populations. The mean age ranged between
NU
61±11 (SD) to 67±9 (SD) years, 73% (1772/2428) were male. The four included studies did not report
MA
on the impact of beta-blockers on HRR separately. In total, 91% (2218/2428) of the study populations were treated with beta-blockers. The proportion of diabetes mellitus was 36% (877/2428). The proportion of hypertension, reported in 3 studies, was 51% (455/897). BMI, reported in 2 studies,
PT E
D
ranged from 28 to 29 kg/m2. The use of calcium channel blockers, digitalis and the prevalence of atrial fibrillation in the studies was reported inconsistently and incompletely and is not reported
CE
here. Study characteristics
AC
Characteristics of the assessment of HRR are presented in table 3. HRR was measured 1 minute postexercise in all 4 studies and the cutoff value for delayed HRR ranged from 12 to 21 beats in the first minute. The Tables 4a and 4b show the statistical methods used in the studies. The adjusted hazard ratio for delayed HRR vs not delayed HRR was reported in all four studies, whereas the unadjusted hazard ratio was reported in three out of four studies. The nature and number of covariates in the various models varied greatly. Detailed information on covariates which were entered into the statistical model is presented in table 4b. 7
ACCEPTED MANUSCRIPT Primary outcome The unadjusted hazard ratio and the pooled unadjusted hazard ratio for the risk of all-cause mortality and composite outcome (cardiovascular death and hospitalization for congestive heart failure), was 5.78 (95%CI 3.20-10.43) (figure 2). I-squared was 32% (p= 0.23), indicating moderate, but statistically non-significant heterogeneity. Table 5 shows the results of the risk of bias assessment of the
PT
included studies; two studies were rated as high quality (22, 23) and two studies were rated as
RI
moderate quality (7, 18).
SC
Discussion
Our meta-analysis, using data from 2484 CAD patients (151 deaths) shows that delayed HRR after
NU
exercise is associated with an increase in all-cause mortality in this patient population. The
MA
incorporation of HRR in routine exercise testing may improve the monitoring of exercise tolerance and risk stratification for mortality in CAD patients.
In general, modifiable cardiovascular risk factors such as overweight, smoking and
PT E
D
cholesterol status are objectively measured using BMI, cotinine testing, and plasma lipid levels. As HRR is more easily measured than other autonomic indices and exercise parameters (18), HRR may be useful as a quick and objective prognostic tool. Particularly, HRR from submaximal exercise test
CE
could potentially be used in monitoring physical fitness. Of the studies included in our meta-analysis, Nissinen et al. reported that HRR is a good
AC
predictor of death compared to traditional autonomic markers such as heart rate variability and baroreflex sensitivity (18). However, Kivinimie et al. found a similar association between HRR and cardiovascular death and hospitalization and maximal chronotropic response index, with an adjusted hazard ratio of 6.0 (2.5-14.5) and 4.9 (2.1-11.3), respectively (23). In the current meta-analysis, the pooled hazard ratio for the risk of all-cause, cardiovascular mortality and composite adverse cardiovascular events was 5.8 (3.2-19.4), which supports stratification of CAD patients according to the risk of mortality. Furthermore, our findings were consistent with a large meta-analysis in 8
ACCEPTED MANUSCRIPT apparently healthy individuals, where delayed HRR was associated with an increased risk of cardiovascular events and all-cause mortality(17). Qiu et al. support the use of HRR for risk assessment in clinical practice. The majority of CAD patients are treated with beta-blockers, potentially influencing HRR. Desai et al. reported that patients treated with beta-blockers have slower HRR compared to patients
PT
without beta-blockers, but they did not report the impact on hazard ratios (24). Considering the pooled hazard ratio of 5.8 in our study, the prognostic value of the HRR for CAD patients seems to
RI
stay intact, notwithstanding the fact that 91% of all patients in our analysis were on beta-blocker
SC
treatment.
NU
The underlying mechanism of impaired HRR
The underlying mechanism of impaired autonomic regulation after peak exercise is not
MA
completely understood. Delayed reactivation of parasympathetic nervous system after cessation of exercise may be induced by a highly activated sympathetic nervous system. Biological links with
D
mortality include impaired vagal reactivation, pro-inflammatory activation, decreased endothelial
PT E
function and arterial stiffness (25-27). Additionally, the prognostic value of HRR has also been observed in other diseases with risk of impaired autonomic responses, such as heart failure, diabetes
CE
mellitus and cancer (27-31). These diseases have in common that stress and inflammatory factors may impair the autonomic function. These stress and inflammatory factors may have a potential role
AC
in the underlying mechanism of impaired HRR in our study population as well. Recovery period and exercise intensity levels HRR has been shown to be prognostic at 1, 2, and 5 minutes after exercise in different populations referred for exercise testing (10, 11, 29, 31, 32). Cole et al. demonstrated that the cutoff value of 12 beats in the first minute of recovery provides maximal discrimination as a prognostic marker for mortality (11), which suggests that parasympathetic reactivity may be more predictive of mortality than sympathetic withdrawal during the later stage of recovery. According to our findings, 9
ACCEPTED MANUSCRIPT the cutoff value of HRR may range from 12 to 21 beats in the first minute of recovery. The cutoff value of 21 beats with a high hazard ratio by Kiviniemi et al. compared to the other included studies in the meta-analysis, may be explained by the difference in outcome measurement, particularly, the use of a composite outcome of mortality and hospitalization of which the latter was not included in the other studies.
PT
In our analysis HRR was derived from symptom-limited cardiopulmonary exercise testing and alternatives such as submaximal exercise test should be considered. In a cohort of 5,234 apparently
RI
healthy individuals, the Lipid Research Clinics Prevalence Study demonstrated that delayed HRR
SC
derived from submaximal exercise testing is significantly associated with mortality (33). Therefore,
NU
submaximal exercise tests might be also useful in monitoring risks and physical fitness at physician check-ups or during cardiac rehabilitation. However, submaximal tests in CAD patients still need to
MA
be validated. Strength and limitations
D
The main strength of this systematic review and meta-analysis was the systematic approach,
PT E
literature search, systematic grading of studies and data extraction. We systematically assessed the statistical analyses in all studies and the prognostic value of HRR was broadly estimated. Limitations
CE
should be noted as well. First, heterogeneity did not allow pooling of individual-level patient data. Second, in one of the studies, the unadjusted hazard ratio was not available even after contacting the
AC
authors (22). Third, there was heterogeneity in co-variates, study outcome, exercise and recovery protocols such as the application of either treadmill testing or the cycle ergometer to measure HRR in the included studies. Since there was heterogeneity in co-variates in the prediction models, we used the unadjusted hazard ratio in the meta-analysis. Furthermore, the heterogeneity of study outcome was due to a composite outcome in one study, which was based on cardiovascular death and hospitalization for congestive heart failure (23). However, all four studies studied mortality as an outcome, and in 2 out of 4 all-cause mortality was studied. And, although there was heterogeneity in 10
ACCEPTED MANUSCRIPT exercise protocols which may lead to differences in maximal workloads, this may not be clinically relevant as the value of HRR as a prognostic marker has been reported to be independent of workload (11, 33). Fourth, chronotropic incompetence and low exercise tolerance, which have also been shown to be associated with the risk of mortality (34, 35), may bias the association between HRR and the risk of mortality. However, according to Kiviniemie et al. HRR independently predicts all-
PT
cause mortality even in the presence of the aforementioned factors (23).
RI
Future research should focus on the added value of HRR in clinical prediction models in large
SC
CAD populations compared to exercise parameters from the Duke treadmill score. Prediction models should have good discrimination of event and no event and calibration (predicted probabilities
NU
should match those observed). Furthermore, since HRR is often derived from symptom-limited exercise testing in CAD patients, the value of HRR derived from submaximal exercise testing as a
MA
metric of physical fitness and an independent prognostic marker in CAD patients needs to be established.
PT E
D
Conclusion
In CAD patients delayed HRR is significantly associated with overall mortality, and provides clinically additional information in this patient population. As HRR is a noninvasive measure of autonomic
CE
function, HRR can be considered in the monitoring of exercise tolerance. Still, future studies are needed to clarify its applicability in non-laboratory settings, in submaximal exercise tests and in
AC
current risk scores.
Acknowledgements
No funders had a role in the study design, data collection, analysis, decision to publish or preparation of the manuscript. Funding sources None 11
ACCEPTED MANUSCRIPT Author contributions SL, MST and RP contributed to the conception and design of the work. SL, MST, JL and GT contributed to the acquisition, analysis and interpretation of data for the work. SL, and MST drafted the manuscript. JL, HT, WSR, CL, SMB, GT and RP critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of work ensuring integrity and accuracy.
PT
Disclosures
RI
SMB reported personal fees from Pfizer, outside the submitted work; RJG reported personal fees
SC
from Sanofi Aventis, Boehringer Ingelheim, Amgen and AstraZeneca, outside the submitted work.
AC
CE
PT E
D
MA
NU
The other authors declare that they have no known conflicts of interest.
12
ACCEPTED MANUSCRIPT References:
AC
CE
PT E
D
MA
NU
SC
RI
PT
1. Sixty-Sixth World Health Assembly. Follow-up to the Political Declaration of the High-level Meeting of the General Assembly on the Prevention and Control of Non-communicable Diseases. WHA66.10, 27 May 2013. http://apps.who.int/gb/ebwha/pdf_files/WHA66/A66_R10-en.pdf. 2. Bansilal S, Castellano JM, Fuster V. Global burden of CVD: focus on secondary prevention of cardiovascular disease. Int J Cardiol. 2015;201 Suppl 1:S1-7. 3. Kotseva K, Wood D, De Bacquer D, De Backer G, Ryden L, Jennings C, et al. EUROASPIRE IV: A European Society of Cardiology survey on the lifestyle, risk factor and therapeutic management of coronary patients from 24 European countries. Eur J Prev Cardiol. 2015. 4. Lee IM, Shiroma EJ, Lobelo F, Puska P, Blair SN, Katzmarzyk PT, et al. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet. 2012;380(9838):219-29. 5. Teo K, Lear S, Islam S, Mony P, Dehghan M, Li W, et al. Prevalence of a healthy lifestyle among individuals with cardiovascular disease in high-, middle- and low-income countries: The Prospective Urban Rural Epidemiology (PURE) study. JAMA. 2013;309(15):1613-21. 6. Anderson L, Oldridge N, Thompson DR, Zwisler AD, Rees K, Martin N, et al. Exercise-Based Cardiac Rehabilitation for Coronary Heart Disease: Cochrane Systematic Review and Meta-Analysis. J Am Coll Cardiol. 2016;67(1):1-12. 7. Hai JJ, Siu CW, Ho HH, Li SW, Lee S, Tse HF. Relationship between changes in heart rate recovery after cardiac rehabilitation on cardiovascular mortality in patients with myocardial infarction. Heart Rhythm. 2010;7(7):929-36. 8. Lucini D, Milani RV, Costantino G, Lavie CJ, Porta A, Pagani M. Effects of cardiac rehabilitation and exercise training on autonomic regulation in patients with coronary artery disease. Am Heart J. 2002;143(6):977-83. 9. Malfatto G, Facchini M, Sala L, Branzi G, Bragato R, Leonetti G. Relationship between baseline sympatho-vagal balance and the autonomic response to cardiac rehabilitation after a first uncomplicated myocardial infarction. Ital Heart J. 2000;1(3):226-32. 10. Ribisl PM, Gaussoin SA, Lang W, Bahnson J, Connelly SA, Horton ES, et al. Lifestyle intervention improves heart rate recovery from exercise in adults with type 2 diabetes: results from the Look AHEAD study. J Obes. 2012;2012:309196. 11. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate recovery immediately after exercise as a predictor of mortality. NEnglJMed. 1999;341(18):1351-7. 12. Girotra S, Keelan M, Weinstein AR, Mittleman MA, Mukamal KJ. Relation of heart rate response to exercise with prognosis and atherosclerotic progression after coronary artery bypass grafting. Am J Cardiol. 2009;103(10):1386-90. 13. Ho JS, Fitzgerald SJ, Barlow CE, Cannaday JJ, Kohl HW, 3rd, Haskell WL, et al. Risk of mortality increases with increasing number of abnormal non-ST parameters recorded during exercise testing. Eur J Cardiovasc Prev Rehabil. 2010;17(4):462-8. 14. Lauer MS, Okin PM, Larson MG, Evans JC, Levy D. Impaired heart rate response to graded exercise. Prognostic implications of chronotropic incompetence in the Framingham Heart Study. Circulation. 1996;93(8):1520-6. 15. Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and sudden cardiac death. Experimental basis and clinical observations for post-myocardial infarction risk stratification. Circulation. 1992;85(1 Suppl):I77-91. 16. Shetler K, Marcus R, Froelicher VF, Vora S, Kalisetti D, Prakash M, et al. Heart rate recovery: validation and methodologic issues. J Am Coll Cardiol. 2001;38(7):1980-7. 17. Qiu S, Cai X, Sun Z, Li L, Zuegel M, Steinacker JM, et al. Heart Rate Recovery and Risk of Cardiovascular Events and All-Cause Mortality: A Meta-Analysis of Prospective Cohort Studies. Journal of the American Heart Association. 2017;6(5). 13
ACCEPTED MANUSCRIPT
AC
CE
PT E
D
MA
NU
SC
RI
PT
18. Nissinen SI, Makikallio TH, Seppanen T, Tapanainen JM, Salo M, Tulppo MP, et al. Heart rate recovery after exercise as a predictor of mortality among survivors of acute myocardial infarction. Am J Cardiol. 2003;91(6):711-4. 19. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6(7):e1000100. 20. Hayden JA, van der Windt DA, Cartwright JL, Cote P, Bombardier C. Assessing bias in studies of prognostic factors. Ann Intern Med. 2013;158(4):280-6. 21. Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594. 22. Aijaz B, Squires RW, Thomas RJ, Johnson BD, Allison TG. Predictive value of heart rate recovery and peak oxygen consumption for long-term mortality in patients with coronary heart disease. Am J Cardiol. 2009;103(12):1641-6. 23. Kiviniemi AM, Lepojarvi S, Kentta TV, Junttila MJ, Perkiomaki JS, Piira OP, et al. Exercise capacity and heart rate responses to exercise as predictors of short-term outcome among patients with stable coronary artery disease. Am J Cardiol. 2015;116(10):1495-501. 24. Desai MY, De la Pena-Almaguer E, Mannting F. Abnormal heart rate recovery after exercise as a reflection of an abnormal chronotropic response. Am J Cardiol. 2001;87(10):1164-9. 25. Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405(6785):458-62. 26. Jae SY, Heffernan KS, Park SH, Jung SH, Yoon ES, Kim EJ, et al. Does an acute inflammatory response temporarily attenuate parasympathetic reactivation? Clin Auton Res. 2010;20(4):229-33. 27. Youn JC, Lee HS, Choi SW, Han SW, Ryu KH, Shin EC, et al. Post-Exercise Heart Rate Recovery Independently Predicts Clinical Outcome in Patients with Acute Decompensated Heart Failure. PLoS One. 2016;11(5):e0154534. 28. Cheng YJ, Lauer MS, Earnest CP, Church TS, Kampert JB, Gibbons LW, et al. Heart rate recovery following maximal exercise testing as a predictor of cardiovascular disease and all-cause mortality in men with diabetes. Diabetes Care. 2003;26(7):2052-7. 29. Giallauria F, Maresca L, Vitelli A, Santucci de Magistris M, Chiodini P, Mattiello A, et al. Exercise training improves heart rate recovery in women with breast cancer. Springerplus. 2015;4:388. 30. Ha D, Choi H, Zell K, Raymond DP, Stephans K, Wang XF, et al. Association of impaired heart rate recovery with cardiopulmonary complications after lung cancer resection surgery. J Thorac Cardiovasc Surg. 2015;149(4):1168-73 e3. 31. Minai OA, Nguyen Q, Mummadi S, Walker E, McCarthy K, Dweik RA. Heart rate recovery is an important predictor of outcomes in patients with connective tissue disease-associated pulmonary hypertension. Pulm Circ. 2015;5(3):565-76. 32. Goda A, Koike A, Iwamoto MH, Nagayama O, Yamaguchi K, Tajima A, et al. Prognostic value of heart rate profiles during cardiopulmonary exercise testing in patients with cardiac disease. Int Heart J. 2009;50(1):59-71. 33. Cole CR, Foody JM, Blackstone EH, Lauer MS. Heart rate recovery after submaximal exercise testing as a predictor of mortality in a cardiovascularly healthy cohort. AnnInternMed. 2000;132(7):552-5. 34. Jouven X, Empana JP, Schwartz PJ, Desnos M, Courbon D, Ducimetiere P. Heart-rate profile during exercise as a predictor of sudden death. N Engl J Med. 2005;352(19):1951-8. 35. Savonen KP, Kiviniemi V, Laukkanen JA, Lakka TA, Rauramaa TH, Salonen JT, et al. Chronotropic incompetence and mortality in middle-aged men with known or suspected coronary heart disease. Eur Heart J. 2008;29(15):1896-902. 14
ACCEPTED MANUSCRIPT Table 1. Characteristics of included studies Kiviniemi et al. 2015 Finland
Country
Hai et al., 2010 China Cardiac rehabilitation
Study Source
Aijaz et al., 2009 USA Cardiac rehabilitation
Nissinen et al., 2003 Finland
Tot al
NU
SC
RI
PT
ARTEMIS Not documented Participants, n 242 (%) 1531 386 282 229 8 Follow-up, years 2.0 6.5 ± 3.4 9.8 ± 2.9 2.6 ± 1.1 Primary Composite All-cause All -cause outcome outcome* Cardiac death* mortality mortality Total events*, n (%) 39 (2.5) 40 (10.4) 55 (19.5) 17 (7.4) 151 Average number of events per year 20 6 6 7 ARTEMIS=Innovation to Reduce Cardiovascular Complications of Diabetes at the Intersection * composite endpoint indicates cardiovascular death and hospitalization for congestive heart failure * cardiac death indicates death due to myocardial infarction, heart failure, or malignant arrhythmia * Total events indicate all-cause, cardiovascular mortality and composite adverse cardiovascular events
Male, n (%)
Resting heart rate (mean±SD)
Hypertension, n (%)
AC
Diabetes mellitus, n (%) Body mass index (mean±SD)
D
Hai et al., 2010 386 (100) 65.2 ± 11.5
Aijaz et al., 2009 282(100) 61.0 ± 11.0
Nissinen et al., 2003 229 (100) 61.5 ± 9.5
1054 (69)
297 (77)
234 (83)
187 (82)
1531 (100)
304 (79)
154 (55)
229 (100)
63.0 ± 12.0
72.0 ± 14.0
71.0 ± 13.0
N/A
CE
Beta-blockers, n (%)
PT E
CAD, n (%) Age (years, mean±SD)
Kiviniemi et al. 2015 1531 (100) 67.0 ± 8.7
MA
Table 2. Baseline characteristics of study participants
N/A
208 (54)
143 (5)
104 (45)
660 (43)
147 (38)
36 (13)
34 (15)
29.0 ± 4.7
N/A
27.6 ± 4.3
N/A
Total (%) 2428 (100) 1.772/2428 (73) 2.218/2428 (91)
455/897 (51) 877/2428 (36)
340/2199 134 (9) 148 (38) 58 (21) N/A (15) LVEF (mean±SD) 62.0 ± 11 45.4 ± 10.0 52.0 ± 13 46.0 ± 9.0 CAD=coronary artery disease, LVEF=left ventricular ejection fraction, N/A=not available and SD=standard deviation. Smoking, n (%)
Table 3. Measurement characteristics of heart rate recovery 15
ACCEPTED MANUSCRIPT
Study
Exercise
Measurement position
Recovery
Recovery (min)
HRR cutoff
Kiviniemi et al., 2015 BET Supine/standing Passive 1 21 Hai et al., 2010 TET Seated Passive 1 12 Aijaz et al., 2009 TET N/A Active 1 13 Nissinen et al., 2003 BET Seated Passive 1 12 All exercise tests were based on maximal level of intensity. BET=bicycle exercise test, TET=treadmill exercise test and HRR=heart rate recovery. The passive recovery means absence of cooling-down after cessation of exercise testing.
Statistical analysis
HRR cutoff
unadjusted HR (95% CI)
adjusted HR (95% CI)
SC
RI
Study outcome Composite U/M Cox ph 21 9.9 (4.4-22.5) 6.0 (2.5-14.5) outcome* U/M Cox ph 12 4.0 (2.1-7.5) 2.5 (1.1-5.6) Cardiac death* All-cause Aijaz et al., 2009 M Cox ph 13 N/A 2.2 (1.1-4.1) mortality Nissinen et al., All-cause 2003 U/M Cox ph 12 5.3 (1.5-18.5) 4.3 (1.2-16.0) mortality U/M Cox ph=univariable or multivariable Cox proportional hazards model, CI=confidence interval, HRR=heart rate recovery, HR=hazard ratio, * composite outcome indicates cardiovascular death and hospitalization for congestive heart failure * cardiac death indicates death due to myocardial infarction, intractable heart failure, or malignant arrhythmia.
AC
CE
PT E
D
MA
NU
Study Kiviniemi et al., 2015 Hai et al., 2010
PT
Table 4a. Statistical analysis
16
ACCEPTED MANUSCRIPT Table 4b. Statistical analysis Study Name Number of covariates in final model Covariates Kiviniemi et al., 2015 10 Age, sex, diabetes mellitus, smoking status, BMI, history of acute myocardial infarction, history of coronary intervention, Canadian Cardiovascular Society grading of angina pectoris, LVEF and resting heart rate. 14
Aijaz et al., 2009
3
Diabetes mellitus, phase 1 LVEF ≤30%, phase 2 LVEF≤30%, acute pulmonary edema, revascularizations before enrollment, use of angiotensinconverting enzyme inhibitor/angiotensin receptor blocker, statin, exercise training, exercise parameters including phase 1 resting heart rate≥65 bpm, phase 1 heart rate increment ≤45bpm, phase 1 exercise capacity, phase duke treadmill score, phase 2 heart rate increment≤45bpm, phase 2 HRR < 12 bpm and phase 2 exercise capacity ≤4 METs.
MA
NU
SC
RI
PT
Hai et al., 2010
Age, sex, low peak VO2.
PT E
D
3 Nissinen et al., 2003 Age, diabetes mellitus and NYHA class BMI=body mass index, LVEF=left ventricular ejection fraction, HRR=heart rate recovery, METs=metabolic equivalents, peak VO2=peak oxygen consumption and NYHA indicates New York Heart Association. Table 5. Quality Assessment based on QUIPS defined by Cochrane Kiviniemi et al., 2015 low moderate low low low
CE
Domains
AC
Study Participants Study Attrition Prognostic factor Outcome Statistical analysis and reporting Overall Score
Hai et al., 2010 Aijaz et al., 2009 moderate low moderate low low low low low moderate moderate
Nissinen et al., 2003 moderate moderate low Moderate low
High quality
Moderate High quality Moderate quality quality QUIPS=Quality in Prognostic Studies tool, high quality ≥ 4 criteria low risk of bias and no criteria of high risk of bias, moderate quality ≥ 2 criteria low risk of bias and the remaining criteria were rated as moderate risk of bias. Low quality ≥ 1 criteria high risk of bias.
17
Figure 1
Figure 2