Article Text

Original research
Comprehensive strain assessment and mortality after acute myocardial infarction: a retrospective observational study based on the Essen Coronary Artery Disease registry
  1. Viktoria Backmann,
  2. Iryna Dykun,
  3. Jürgen Kampf,
  4. Bastian Balcer,
  5. Anja Roggel,
  6. Matthias Totzeck,
  7. Tienush Rassaf,
  8. Amir A Mahabadi
  1. Department of Cardiology and Vascular Medicine, University Hospital Essen, Essen, Germany
  1. Correspondence to Amir A Mahabadi; Amir-Abbas.Mahabadi{at}uk-essen.de

Abstract

Background and aims Current ESC guidelines on the management of patients after acute myocardial infarction only include the evaluation of left ventricular (LV) function by assessment of the ejection fraction in addition to clinical risk scores to estimate the patient’s prognosis. We aimed to determine, whether comprehensive evaluation of cardiac function using LV and right ventricular (RV) global longitudinal strain (GLS) and left atrial (LA) reservoir strain improves the prediction of survival in patients with acute myocardial infarction.

Methods In patients with non-ST segment elevation or ST segment elevation myocardial infarction receiving echocardiography within 1 year after revascularisation, LV-GLS, RV-GLS and LA reservoir strain were quantified. In multivariable Cox regression analysis, HRs and 95% CIs were calculated per 1 SD increase in strain measure, adjusting for age, sex, systolic blood pressure, low-density lipoprotein cholesterol, smoking, diabetes and family history of premature coronary artery disease.

Results During a median follow-up of 1.5 (0.5–4.2) years, 157 (11.1%) out of 1409 patients (64.4±13.5 years, 24.7% female) died. LV-GLS (1.68 (1.37–2.06), p<0.001), RV-GLS (1.39 (1.16–1.67), p<0.001) and LA reservoir strain (0.57 (0.47–0.69), p<0.001) were associated with mortality. Adding LV ejection fraction, tricuspid annular plane systolic excursion (TAPSE) or LA volume index to these models did not alter the association of strain measures of the LV (1.41 (1.06–1.89), p=0.02), RV (1.48 (1.03–2.13), p=0.04) or LA (0.61 (0.49–0.76), p<0.001). In receiver operating characteristics, combining the three strain measures improved the prediction of mortality above risk factors (AUC: 0.67 (0.63–0.71) to 0.75 (0.70–0.80)), while further addition of LV ejection fraction, TAPSE and LA volume index did not (0.75 (0.70–0.81)).

Conclusion The comprehensive evaluation of contractility of various cardiac chambers via transthoracic echocardiography using myocardial strain analysis, when routinely performed after acute myocardial infarction, may help to detect patients at increased mortality risk.

  • acute myocardial infarction
  • echocardiography

Data availability statement

Data are available upon reasonable request. The data underlying this article will be shared on reasonable request to the corresponding author.

http://creativecommons.org/licenses/by-nc/4.0/

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

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Previous studies suggested that echocardiographic parameters other than left ventricular (LV) ejection fraction like global longitudinal strain (GLS) and left atrial (LA) reservoir strain are useful for risk stratification and assessment of prognosis after acute myocardial infarction.

  • However, existing literature, focused on specific patient cohorts, were limited by a low number of events or were based on strain assessment of a single cardiac chamber.

WHAT THIS STUDY ADDS

  • To the best of our knowledge, this is the first large cohort of consecutive patients with acute myocardial infarction—both non-ST segment elevation and ST segment elevation myocardial infarction—evaluating the independent predictive value of LV-GLS, right ventricular GLS and LA reservoir strain within the same cohort.

  • We report strong associations of each strain measure on adjustment for cardiovascular risk factors and traditional echocardiographic measures with strain measurements significantly improving the prediction of incident mortality.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE AND POLICY

  • Our results suggest that adding strain measurements to the routine evaluation in patients following acute myocardial infarction could improve the detection of high-risk cohorts leading to different clinical pathways, treatment decisions and follow-up visits.

Introduction

Despite improvement in management of patients with myocardial infarction, long-term mortality remains high.1 For improved patient management, detecting high-risk patients is essential.2 3 Echocardiography is a low-cost and broadly available imaging technology, enabling healthcare professionals to non-invasively assess cardiac anatomy and function.4 Current European Society of Cardiology guidelines on the management of patients after acute myocardial infarction include the assessment of left ventricular (LV) function by quantification of the ejection fraction in addition to clinical risk scores to estimate the prognosis.2 3 However, many patients face an unfavourable prognosis after acute myocardial infarction despite a normal or only moderately reduced LV ejection fraction.5–7 Previous studies suggested that other echocardiographic parameters like tricuspid annular plane systolic excursion (TAPSE), and left atrial (LA) volume index, as well as global longitudinal strain (GLS) and LA reservoir strain are useful for risk stratification after acute myocardial infarction.8–15 However, existing literature were limited by specific patient cohorts,9 a low number of events5 or were based on strain assessment of a single cardiac chamber.5 9 In the present analysis, we aimed to determine, whether comprehensive strain assessment of LV, right ventricular (RV) and LA function associates with survival in patients after myocardial infarction in a large cohort of consecutive patients. Specifically, we aimed to determine whether this would improve prediction of mortality, independent of traditional cardiovascular risk factors and conventional echocardiographic measures.

Methods

Study cohort

The present analysis is based on a retrospective registry of consecutive patients undergoing invasive coronary angiography at the West German Heart and Vascular Center, Essen, between 2004 and 2019 (the Essen Coronary Artery Disease (ECAD) registry), including data from 40 461 coronary procedures (data set as of July 2019). Details regarding the ECAD registry have been published elsewhere.16

For this analysis we included patients with the primary discharge diagnosis of non-ST segment elevation or ST segment elevation myocardial infarction and available transthoracic echocardiography imaging within 1 year after index myocardial infarction of sufficient quality for strain assessment.

Clinical characteristics and endpoint definition

Information on traditional cardiovascular risk factors from the same hospital stay was obtained from the hospital information system. The primary endpoint variable was all-cause mortality. Details for assessment of clinical characteristics and echocardiographic measurements are described in the online supplemental file 1.

Supplemental material

Strain analysis

A single experienced reader, blinded to the clinical presentation and outcome of the patients, assessed the strain measures offline at a central core lab using the TOMTEC- Arena 2D Cardiac Performance Analysis software (Philips Healthcare, Best, The Netherlands). Details are found in the online supplemental file 1.

Statistical analysis

The baseline characteristics are presented as the mean±SD for normally distributed continuous variables, as median and IQR for non-normally distributed continuous variables, and as frequencies and percentages for categorical variables. Two-sided t-tests were used for normally distributed continuous variables, Mann-Whitney U tests for non-normally distributed continuous variables and χ2 tests for categorical variables for comparison of survivors and non-survivors. Cox regression analysis was used to determine the association of strain measures with incident mortality. Adjustment sets were defined as follows for each strain measure: (1) Unadjusted; (2) Adjusted for age, sex, systolic blood pressure, low-density lipoprotein-cholesterol, smoking status, diabetes and family history of premature coronary artery disease; and (3) Ancillary models with the addition of LV ejection fraction or TAPSE or LA volume index. HRs and 95% CIs are depicted per 1 SD increase in strain measure. For the multivariable regression analysis, multiple imputation of covariates was performed in case of missing data creating five imputed data sets. The results obtained from each completed imputed data set were then combined into one multiple-imputation result. A sensitivity analysis without multiple imputation was performed, as well as one excluding patients with bypass surgery and one adjusting for significant valvular heart disease. In another sensitivity analysis the cohort was divided into subgroups of ≥ versus < median of LV ejection fraction, creatine kinase, and troponin levels and Cox regression analysis was performed in each of these fully adjusted multivariable models. Furthermore, subgroup analysis was also performed for patients with non-ST segment elevation versus ST segment elevation myocardial infarction. Receiver operating characteristics analysis was performed to evaluate the predictive value of the strain measures using three models defined as: (1) Age, sex, systolic blood pressure, low-density lipoprotein-cholesterol, smoking status, diabetes, family history of premature cardiovascular disease, (2) Model 1 plus LV-GLS, RV-GLS and LA reservoir strain; and (3) Model 2 plus LV ejection fraction, TAPSE and LA volume index. The DeLong method was used to compare the area under the receiver operating characteristics curves. Furthermore, Cox regression analysis was conducted to determine the association of each strain measure with incident mortality additive to the Global Registry of Acute Coronary Events (GRACE) Score. All analyses were performed using SPSS software (V.28.0.0.0 (190), IBM SPSS Statistics), except for the receiver operating characteristics comparison, which was performed using MedCalc (V.20.113). A two-sided value of p<0.05 indicated statistical significance.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

Results

Figure 1 depicts the study flow chart. Cases without follow-up information (n=6483), as well as patients with chronic coronary syndrome (n=20 473), unstable angina pectoris (n=3631), non-coronary interventions (n=3117) and non-cardiac diagnoses (n=4319) were excluded. From the remaining cohort, postinterventional transthoracic echocardiography with sufficient image quality was not available in 1029 cases, leaving an overall cohort of 1409 patients (mean age 64.4 years, 75.3% male). As not all echocardiographic images allowed the measuring of all three strain parameters, we were able to include 1078 patients with LV-GLS, 1073 with RV-GLS and 1369 with LA reservoir strain. 970 patients (68.8%) were diagnosed with non-ST segment elevation myocardial infarction, 439 patients (31.2%) with ST segment elevation myocardial infarction. The median time between index procedure and echocardiography was 4 (2–13) days. During a median follow-up of 1.5 (0.5–4.2) years, 157 deaths (11.1%) occurred. Table 1 depicts detailed patient characteristics.

Table 1

Patient characteristics stratified by survivors versus non-survivors

Figure 1

Study flow chart. AMI, acute myocardial infarction; ECAD registry, Essen Coronary Artery Disease registry; IAP, instable angina pectoris; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction.

Patients who died were on average slightly older, more frequently male and more frequently diabetics. They also had a lower LV ejection fraction, lower TAPSE, larger LA volume index and higher RV end-diastolic diameter as well as poorer strain values.

Table 2 depicts the Cox regression analysis for the association of the strain measurements with survival.

Table 2

Cox regression analysis, depicted per 1 SD of strain measure, unadjusted and adjusted for CVRF, LVEF, TAPSE and LAVI

In unadjusted regression analysis, all three strain measures were strongly associated with incident mortality. Effect sizes remained stable on adjustment for traditional cardiovascular risk factors. To evaluate whether strain measurements add value to established measures of cardiac size and function, we added both into one multivariable model. When both LV-GLS and LV ejection fraction were applied into one model, only LV-GLS but not LV ejection fraction (HR (95% CI) 0.78 (0.59–1.03), p=0.07) remained independently associated. Likewise, RV-GLS, but not TAPSE (0.80 (0.56–1.16), p=0.25), independently associated with mortality. As TAPSE can be altered after open heart surgery, we performed a sensitivity analysis excluding patients with bypass surgery (n=197, 14%). In the remaining subgroup (n=1212), TAPSE was slightly higher as compared with the overall cohort. In Cox regression analysis, however, associations of RV-strain and TAPSE in combined models remained unchanged (RV-GLS: 1.54 (1.04–2.30); p=0.032; TAPSE: 0.84 (0.56–1.27); p=0.405).

However, simultaneously adding LA reservoir strain and LA volume index into one fully adjusted multivariable model, both LA reservoir strain and LA volume (1.20 (1.03–1.41), p=0.02) remained significantly associated with mortality. The sensitivity analysis without multiple imputation and the sensitivity analysis adjusting for significant valvular heart disease showed no significant change of results, as shown in online supplemental tables 1 and 2.

Figure 2 describes the fully adjusted multivariable models of the relationship between the strain measures and all-cause mortality, stratified according to subgroups of LV ejection fraction, creatine kinase and troponin levels of < versus ≥ median. Overall, we observed no significant interaction for all three strain measures with any subgroup, while HRs for the association of LV-GLS with mortality were numerically higher in patients with high creatine kinase and high troponin levels and numerically lower for patients with preserved LV ejection fraction.

Figure 2

Association of strain assessment and all-cause mortality stratified according to subgroups of < versus ≥ median LVEF, CK and troponin levels. Cox regression analysis is depicted per 1 SD of strain measure, adjusted for age, sex, systolic blood pressure, low-density lipoprotein cholesterol, smoking status, diabetes, positive family history of premature coronary artery disease. CK, creatine kinase; GLS, global longitudinal strain; LA, left atrium; LV, left ventricle; LVEF, left ventricular ejection fraction; RV, right ventricle.

To further evaluate the prognostic value of these strain measures, we compared the effect sizes for non-ST segment elevation and ST segment elevation myocardial infarction patients. As shown in online supplemental table 3, patients diagnosed with non-ST segment elevation myocardial infarction were older, had a higher body mass index, higher systolic blood pressure, were more likely to be diabetics and less likely to smoke compared with patients diagnosed with ST segment elevation myocardial infarction. Additionally, they had lower TAPSE, higher RV end-diastolic diameter and a larger LA volume. Furthermore, while the LV-GLS values did not differ, patients with non-ST segment elevation myocardial infarction also had higher RV-GLS and lower LA reservoir strain. In multivariable Cox regression analysis, the strain measures were all associated with incident mortality in both subgroups, with slightly higher effect sizes for patients with ST segment elevation myocardial infarction (table 3).

Table 3

Association of GLS assessment and all-cause mortality stratified according to subgroups of patients with NSTEMI and STEMI. Cox regression analysis is depicted per 1 SD of strain measure, adjusted for age, sex, systolic blood pressure, low-density lipoprotein cholesterol, smoking status, diabetes and positive family history of premature coronary artery disease

In receiver operating characteristics analysis, combining LV-GLS and RV-GLS, and LA reservoir strain significantly improved predictions of mortality above traditional risk factors (risk factors: 0.67 (0.63–0.71), risk factors+strain: 0.75 (0.70–0.80), p=0.008). Adding LV ejection fraction, TAPSE and LA volume index to the model did not further improve the area under the receiver operating characteristics curve of 0.75 (0.70–0.81), p=0.47 (figure 3).

Figure 3

Receiver operating characteristics curves for the prediction of all-cause mortality for cardiovascular risk factors (blue line), cardiovascular risk factors and LV, RV and LA reservoir strain (green line), as well as when ancillary adding LVEF, TAPSE and LA volume index to the model (orange line). CVRF, cardiovascular risk factors; LA, left atrial; LAVI, left atrial volume index; LV, left ventricular; LVEF, left ventricular ejection fraction; RV, right ventricular; TAPSE, tricuspid annular plane systolic excursion.

The comparison to the GRACE Score showed an additive predictive value for each strain measure as shown in online supplemental table 4.

Discussion

Within the present large longitudinal observational registry on consecutive patients with acute myocardial infarction undergoing percutaneous coronary intervention, we demonstrated that LV-GLS and RV-GLS and LA reservoir strain associate with incident mortality, independent of traditional risk factors. The strong link between strain measures and impaired survival was equally present for patients with non-ST segment elevation and ST segment elevation myocardial infarction as well as other markers of infarct size. Adding LV ejection fraction, TAPSE and LA volume index as conventional echocardiographic markers to the model did not further improve the prediction of incident mortality. Therefore, our results suggest that the comprehensive evaluation of contractility of various cardiac chambers via transthoracic echocardiography using strain analysis may help to detect patients at increased incident mortality risk.

Echocardiography and risk assessment after myocardial infarction

Echocardiography is a low-cost and broadly available imaging technology that non-invasively assesses cardiac anatomy and function.4 Extensive literature documents the value of LV ejection fraction as a prognostic marker after acute myocardial infarction.4 6 8 17 This evidence is reflected by current European Society of Cardiology guidelines regarding non-ST segment elevation and ST segment elevation myocardial infarction, recommending the assessment of LV ejection fraction in addition to clinical risk scores to estimate patients’ prognosis.2 3 However, LV ejection fraction has several limitations, as it is angle dependent and while there is an inverse relationship with mortality, a normal or only moderately reduced LV ejection fraction may be unrelated to mortality.5–7

In comparison to LV ejection fraction, LV-GLS is relatively operator independent and more reproducible because of its semiautomated assessment.4 6 8 18 High intraobserver and interobserver agreement is fundamental to routine clinical adoption of this measurement. In this study a single observer obtained the strain measurements using echocardiographic images recorded by several experienced cardiologists. Previous studies suggested that LV-GLS provides additional prognostic information and is more sensitive to LV dysfunction after acute myocardial infarction than LV ejection fraction.4 6 8 19–21 As subendocardial fibres are more prone to ischaemic damage and thereby altered early after acute myocardial infarction,4 18 19 22 LV-GLS measuring the subendocardial longitudinal shortening may allow to detect LV dysfunction, which ultimately does not lead to impaired ejection fraction. A study with a cohort of 445 first-time patients with ST segment elevation myocardial infarction receiving percutaneous coronary intervention reported LV-GLS measured by cardiac magnetic resonance as an independent predictor for major adverse cardiovascular events superior to LV ejection fraction.21 Likewise, a study on 659 patients with acute myocardial infarction described that LV-GLS associated with all-cause mortality and a composite of non-fatal re-infarction, coronary revascularisation and hospitalisation for heart failure.8 In our large cohort of patients with non-ST segment elevation and ST segment elevation myocardial infarction LV-GLS measured by transthoracic echocardiography was also superior to LV ejection fraction, as LV ejection fraction did not remain significantly associated with mortality when adding LV-GLS to multivariable models.

In addition to LV function, several studies have indicated that RV function is a critical marker of clinical outcomes in patients after acute myocardial infarction.9 17 23 It can be assessed using various imaging modalities. TAPSE measures the systolic excursion of the lateral tricuspid valve annulus and has been reported to predict mortality in patients with acute myocardial infarction.24

A study with a cohort of 682 patients undergoing percutaneous coronary intervention for acute myocardial infarction reported both RV-GLS and TAPSE as strong univariate predictors for adverse outcomes. However, only RV-GLS but not TAPSE provided independent prognostic information in multivariable analysis.17 Park et al reported similar results for 282 patients with inferior ST segment elevation myocardial infarction.9 However, those studies were of smaller size and used an RV-GLS measurement, which required manual delineation and modification. We were able to confirm their findings using an automated, commercially available software package. While in our analysis both RV-GLS and TAPSE were associated with mortality after adjustment for traditional risk factors, only RV-GLS remained significant when adding both into the same model. Park et al also reported that RV-GLS was not associated with adverse clinical outcomes, if patients had LV systolic dysfunction (LV ejection fraction <50%).9 In contrast, in our study GLS measurements did not show any interactions with LV ejection fraction.

Lastly, in patients experiencing acute myocardial infarction, LA function may be directly involved in acute coronary syndrome.25 Furthermore, LV ischaemia affects LA function, leading to increased LA pressure and contraction to compensate for LV dysfunction.26 Additionally, previous studies stated that LA relaxation and reservoir function is of importance during acute ischaemia.26 While LA volume depends on geometrical assumptions and is angle dependent,27 LA reservoir strain allows for a direct evaluation of myocardial deformation. A study with a cohort of 320 patients after acute myocardial infarction found that LA volume index and LA reservoir strain were independently associated with adverse outcome.28 Another study on patients with first ST segment elevation myocardial infarction reported that LA reservoir strain evaluation led to an improved reclassification of major adverse cardiovascular events in addition to E/e’, degree of mitral regurgitation, RV fractional area change and TAPSE.29 We were able to confirm these findings in our study, extending the literature to patients with recurrent acute myocardial infarction and non-ST segment elevation myocardial infarction. In our study, the effect sizes were slightly higher for patients with ST segment elevation as compared with non-ST segment elevation myocardial infarction, while remaining statistically significant in both entities.

In the present analysis, both patients with ST segment elevation as well as non-ST segment elevation myocardial infarction were included. To evaluate, whether the observed associations were predominantly driven by one myocardial infarction entity, we performed a subgroup analysis. Here, we observed comparable effect sizes, suggesting that comprehensive evaluation of cardiac function using strain may equally identify patients at impaired survival when presenting with ST segment elevation as well as non-ST segment elevation myocardial infarction.

To the best of our knowledge, this is the first large cohort of consecutive patients with acute myocardial infarction, evaluating the independent predictive value of LV-GLS and RV-GLS, and LA reservoir strain within the same cohort. We report strong associations of each strain measure on adjustment for cardiovascular risk factors and traditional echocardiographic measures with strain measurements significantly improving the prediction of incident mortality. Therefore, our results suggest the routine comprehensive evaluation of LV, RV and LA function using GLS and reservoir strain via transthoracic echocardiography in patients following acute myocardial infarction.

Clinical application

Clinicians strive towards better risk assessment to identify groups of patients at increased mortality risk that could benefit from intensified treatment. Semiautomated assessment of strain measures using commercially available software is reproducible, relatively operator independent, and can be added to a standard echocardiographic assessment with minimal additional efforts. Therefore, our results suggest that adding strain measurements to the routine evaluation in patients following acute myocardial infarction could improve the detection of high-risk cohorts, leading to different clinical pathways as well as treatment decisions, including medication and follow-up visits.

Study limitations

First, our results are based on a retrospective single-centre experience using administrative information for follow-up assessment. While the database includes cases from a period of 16 years with multiple interventional cardiologists, echocardiographic technicians and physicians as well as different echocardiography machines, our results need to be confirmed in cohorts from other centres and different healthcare systems. Second, only patients who survived coronary angiography, myocardial revascularisation and underwent echocardiography were included in our study, which may represent a selection bias. The exclusion of patients with missing echocardiographic data may have led to an underestimation of the true prognostic value of strain measures. Third, in order to avoid bias due to missing data, we performed multiple imputations. Although no estimation method is fail-safe, multiple imputation is considered the optimal approach regarding missing values.30 Furthermore, our analysis is based on a predominantly Caucasian population; hence, generalisation to other ethnic groups remains uncertain. Lastly, further prospective studies are needed for external validation of our results, to assess the preintervention to postintervention changes of strain measurements and demonstrate that interventions based on these improve outcomes.

Conclusion

In our large longitudinal observational registry on consecutive patients with acute myocardial infarction undergoing percutaneous coronary intervention, strain assessment of LV, RV and LA function improved risk assessment regarding incident mortality independent of traditional risk factors and conventional measures of cardiac size and function. This effect was equally present for patients with non-ST segment elevation as well as ST segment elevation myocardial infarction. Our results suggest that comprehensive evaluation of contractility of various cardiac chambers via transthoracic echocardiography using myocardial strain analysis, when routinely performed after acute myocardial infraction, may allow detecting patients at increased mortality risk.

Data availability statement

Data are available upon reasonable request. The data underlying this article will be shared on reasonable request to the corresponding author.

Ethics statements

Patient consent for publication

Ethics approval

The analysis was approved by the local ethics committee (19–8956-BO).

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

  • Contributors All authors contributed to the design and implementation of the research, to the analysis of the results and to the writing of the manuscript. AAM is the corresponding author and guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • 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.