On March 18th, 2024, we searched PubMed for identifying indexed clinical guidelines from January 1990 until December 2023 published by the ACC, AHA, and ESC. The search string included controlled and noncontrolled terminologies and is shown in Supplementary Appendix 1 (pp. 7-8). Additionally, a supplementary search for relevant guidelines, corresponding to a focal and single-dated consultation on the beforementioned cardiovascular societies’ online platforms as well as the immediate predecessors to any eligible shortlisted guideline, was then carried out. We also performed an exhaustive search of reference lists of eligible guidelines, ensuring the inclusion of the highest number of records. ^,, The systematic search and guidelines selection were performed by an experienced medical researcher, and discrepancies were solved by consultation with a senior cardiologist. Once publications were identified, they were imported onto Covidence for screening (title and abstract and sequentially full texts).

As recommendations pertaining to STEMI management may be found in guidelines not specifically focusing on the management of ACS, we carefully evaluated guideline documents focused on coronary artery revascularization techniques (ie, percutaneous coronary intervention [PCI] and coronary artery bypass graft [CABG] surgery), management of coronary artery disease, and coronary angiography, and emergency cardiovascular care. For the eligible guidelines covering the spectrum of ACS, we excluded recommendations clearly labeled to be applicable to the management of unstable angina and Non–ST-segment Elevation Myocardial Infarction (NSTEMI or non-Q-wave myocardial infarction per old definitions), as well as those related to the management of conditions associated solely with chronic coronary syndromes. Moreover, we excluded recommendations under the category of risk stratification, physical examination, electrocardiogram monitoring, triage and transportation protocols, diagnostic cardiac biomarkers, pre and in-hospital logistic of care, as well as follow-up after the acute event. These exclusions ensure that our focus remains only on those recommendations pertinent to in-hospital STEMI management.

The definition of terminologies utilized in our study followed the endorsements by the World Health Organization (WHO), the ACC, and the AHA. The full determination of all relevant terminologies pertaining to this research (clinical guidelines, recommendations, focused updates, pharmacological interventions, nonpharmacological interventions, and acute STEMI) is available in Supplementary Appendix 2 (pp. 9-10). Eligible guidelines statements were those in which the recommendation clearly focused on an intervention intended for management of STEMI in the acute phase, dichotomized as pharmacological or nonpharmacological approaches. Listed pharmacological recommendations were framed into pharmacological subgroups (third level), according to the Anatomical Therapeutic Chemical (ATC) Classification System, controlled by the World Health Organization Collaborating Centre for Drug Statistics Methodology (WHOCC). nonpharmacological recommendations during the acute STEMI management encompassed therapies that do not essentially and entirely involve taking active chemical compounds (ie, PCI, coronary angiography, ventricular assistance devices, and intra-aortic balloon counter pulsation). As our analytical focus on clinical trials of interventions, we included only recommendations applicable to the in-hospital management of STEMI patients. Thus, we dismissed recommendations related to risk stratification, physical examination, electrocardiogram (ECG) monitoring, triage and transportation protocols, diagnostic cardiac biomarkers, pre and in-hospital logistic of care (ie, emergency medical services, public education in cardiopulmonary resuscitation, and admission to cardiac coronary unit’s procedures), as well as follow-up after the acute event. The comprehensive list of pharmacological and nonpharmacological interventions considered eligible for abstraction is available in Supplementary Appendix 3 (pp. 11-20).

Data extraction was initially performed by 1 methodologist-physician, with subsequential assessment of completeness by a peer-researcher specialized in cardiovascular research. The final list of recommendations was prepared later under the supervision of a senior interventional cardiologist and collaborating with a senior epidemiologist. Any analytical or theoretical uncertainty by the primary clinicians was solved through remote or in-person discussions, as were discrepancies related to the extraction of recommendation data. Besides recording the text of each recommendation, we extracted each recommendation’s COR and LOE. We did not provide any further judgment or assessment of the proposed definition displayed within the guidelines, and we utilized the same classification as originally reported by the guideline. Along with the categorization of the LOE among included guidelines, we extracted the CORs encompassing 5 levels. The full description of all 3 LOE and the 5 COR as described in Supplementary Appendix 2 (pp. 9-10).

For capturing the temporal evolution of recommendations, we tabulated all extracted data from individual guideline documents in a single database. We further randomly assigned codes (identifiers) to the recommendations, using a coding system presented in detail in the Supplementary Appendix 4 (pp. 21-48). The coding system utilized principles of semantic similarity and thematic-based clusterization to identify equivalent recommendations over time. For most recommendations, we followed the clinical guidelines subheading identifiers to perform the matching of equivalent codes across the database, always maintaining the determined recommendation under this major category of original reporting. Recommendations on the various types of cardiac implantable electronic devices (eg, implantable cardioverter defibrillator, ICD) were carefully allocated under a general group (ICD), and subtypes within each category were also labeled (eg, implantable defibrillators, permanent pacing, transvenous catheter pace, temporary pacing, and temporary transvenous pacing). However, for some recommendations (such as those related to management of complications and the use of prophylactics or symptomatic drugs) we decided to assign them to a separate category based on our clinical experience. We managed the overlapping of recommendations on an equivalent code by checking the consistency among guidelines (evaluating each directly related predecessor, in the case of subsequential documents). Nevertheless, we deemed as different those recommendations from the same guideline document that were considered semantically and technically equivalent but with contrasting COR and/or LOE. To capture the temporal evolution of both pharmacological and nonpharmacological recommendations, we initially assessed the annual proportion of class-specific CORs and LOE across the database. We expressed these results as the total number (N) and percentage (%), by year and subcategory. Lastly, we compared the focal changes of LOE from 2 comparable guidelines (eg, ACC/AHA [2004/2013] and ESC [2017/2023], using the chi-square test). We reported out data graphically using heat maps, linear plotting graphs, and Sankey diagram, stressing the changes in CORs and LOEs over time. We processed and tracked all the modifications of any subcategory of recommendation at different LOEs. Two medical statisticians coordinated the mathematical assessments using R software (version 4.3.1).

Our electronic search yielded 805 references (as 775 unique publications), with 136 duplicates automatically identified by our systematic review management platform. On the first stage of screening (title and abstract), we considered 567 studies as irrelevant (Supplementary Appendix 5, pp. 52-110). Throughout full text evaluation, 72 records were shortlisted, with 54 records being excluded (reasons reported in Figure 1 and Supplementary Appendix 6, pp. 111-118). With regard to our hand search, 8 additional guidelines were obtained. Therefore, our meta-research analysis collates data from 26 clinical practice guidelines (published as 49 records) issued by the selected cardiology societies, from 1990 to 2023 (1990, ^, 1995, ^,, 1996, ^, 1999, ^, 1999, ^, 1999, ^, 2000, 2001, ^, 2003, ^, 2003, ^, 2004, ^, 2005, 2005, 2008, ^, 2008, 2009, ^,, 2011, ^,, 2013, ^, 2014, ^, 2015, ^,, 2016, ^, 2016, ^, 2017, 2017, 2021, ^, 2023 ), as showed in Table 1 . Guidelines were predominantly published in the Journal of the American College of Cardiology ( n = 18 ), Circulation ( n = 18 ), and European Heart Journal ( n = 5 ). Although mostly endorsed by the ACC, AHA, or the ESC, some included guidelines had endorsement approval from other medical organizations, including the Society for Cardiovascular Angiography and Interventions (SCAI) ^,,,,, and the American Association for Thoracic Surgery (AATS). ^,,

Review flowchart diagram.

Our historical registration of recommendations from 1990 to 2023 encompasses 2,139 recommendations across 26 guidelines. Table 2 depicts the annual total number of recommendations and their associated LOEs and CORs for the 3 selected cardiology societies. Guidelines covered specific topics or were focused updates (eg, PCI, ^,,,,, coronary artery revascularization strategies, ^,,, and dual antiplatelet therapy ^, ) or were presented as an overarching guidelines (on both ACS, acute myocardial infarction, and STEMI ^,,,,,,,,,, ). For the 33 years evaluated, the LOE fluctuated across the years, with the highest proportion of high LOE (LOE-A) registered in 1999 and 2021. Figure 2 shows the general evolution of the LOE ( Figure 2 A) and COR ( Figure 2 B) throughout the tracked period. After analyzing the dataset to determine if a significant change over the years occurred, we observed a constant fluctuation in the number of recommendations under an assigned high, moderate, and low LOE (significant change in the proportion among the LOEs over the years (χ² = 114.65, P -value <.01). Despite presenting the distribution of LOE across included guidelines through proportion visualization, we acknowledge the fact that they may obscure the absolute increase in LOE-A recommendations in case the total number of recommendations concomitantly grows. Thus, we are also presenting absolute counts and proportions to provide a more balanced view of tendencies across the time ( Table 2 ).

Trends in LOE (A) and COR (B) in STEMI-related guidelines (1999-2023).

Figure 3 illustrates the proportions (as %) of different LOE across the multiple COR. The heatmap suggests that Class I recommendations have a relatively low number of high-quality evidence. Considering all included recommendations, we observed that approximately 70% of highly advised clinical recommendations are based on either moderate (37.3%) or low-quality evidence (32.6%). Similarly, Class II-a and Class II-b recommendations were shown to have a stronger reliance on lower-quality evidence, precisely 45.9% and 49.3%, respectively. As far as Class III recommendations are concerned, they predominantly relied on moderate-quality evidence (proportion of 45.1%). We further evaluated the variation and concentration of different LOE over the years for different pharmacological and nonpharmacological categories. Figures 4A, 4B, and 4C (Supplementary Appendix 7, pp. 120-122) show the temporal evolution of the LOE among all categorized pharmacological interventions (LOE-A, LOE-B, and LOE-C, respectively) while in Figures 5A, 5B, and 5C (Supplementary Appendix 7, pp. 123-125) we report the different LOE among all nonpharmacological interventions. There was substantial variability by year in COR and LOE across different pharmacological and nonpharmacological classes.

Cross-tabulation ofLOE and COR in STEMI-related guidelines (1999-2023).

The ACC/AHA, and ESC started to systematically report the LOE for each recommendation within their clinical practice guidelines after 2001 and 2003, respectively. With regards to the ACC/AHA guidelines, from 2001 to 2021, most recommendations were based on both LOE-B ( n = 360; 45.6%) and LOE-C ( n = 309; 39.2%), with only 15.1% of recommendations being classified as LOE-A ( n = 119 ). Similarly, the ESC evidenced a slightly similar pattern, with most recommendations based on low LOE ( n = 319; 37.7%) and moderate LOE ( n = 293, 32.9%), and a smaller percentage of recommendations based on high LOE ( n = 278; 32.8%). To note, our assessment comparing the LOE between the ACC/AHA and the ESC revealed a statistically significant difference in the distribution of LOE across the 3 cardiology societies (χ² = 54.118, df = 4, P -value < 0.001). The generated Sankey diagram (Figure 6, Supplementary Appendix 7, pp. 123) illustrates the temporal changes in the LOEs of included recommendations from multiple years, suggesting major patterns. Recommendations from earlier years (ie, 2004 and 2008), are predominantly under a moderate- and low-quality of evidence. Additionally, we observe the persistence of recommendations classified as LOE-C across different years, stressing the recurring trend of recommendations with low-quality of evidence. Lastly, it is observed a high variability in the strength of evidence underpinning included recommendations over time.

The sub categorical assessment evidenced that pharmacological interventions presented a slightly higher proportion of recommendations under the high LOE (21.5%), compared to nonpharmacological interventions (13.8%), P -value < 0.01. nonpharmacological interventions have shown to have slightly higher proportion of recommendations under a low LOE compared to pharmacological interventions (30.4% vs 29.4%, P -value <.05). To note, both types of cardiovascular-related interventions predominantly were associated with Class I recommendations, in which nonpharmacological interventions presenting a greater proportion of recommendations under this category compared to pharmacological interventions (73.5% vs 68.8%). Interestingly, nonpharmacological recommendations presented a slightly larger proportion of Class III recommendations (9.9%) compared to pharmacological interventions (7.1%). Our statistical analysis also suggested that significant changes in pharmacological versus nonpharmacological recommendations occurred over time (χ² = 267.25, P -value < 0.05), as well as in the subareas within each type of recommendations (pharmacological subareas yielded a χ² = 1,159.5; P -value < 0.05, while nonpharmacological subareas yielded a χ² = 1,592.16; P -value < 0.05). The full list of all subcategories of pharmacological and nonpharmacological interventions included in our analysis is presented in Tables 3 and 4 . Due to space limitations in the main publication, we included additional findings summarization in Supplementary Appendix 7 (pp. 119-125).

Table 3

Distribution of the level of evidence (LOE) and class of recommendation (COR) of pharmacological interventions, by subcategory.

Subcategory of pharmacological interventions	Pharmacological COR				Pharmacological LOE
	I	II-a	II-b	III	A	B	C
Agents acting on the renin-angiotensin-system (C09)	18	9	0	1	24	4	0
Aldosterone antagonists and other potassium-sparing agents (Mineralocorticoid receptor antagonist, C03D)	13	0	0	0	6	7	0
Analgesic (N02), other analgesics and antipyretics (N02B), or ant-inflammatory agents (S01B)	8	4	6	5	0	4	15
Angiotensin II receptor blockers (ARBs), Plain (C09C)	17	7	3	0	4	10	3
Antiarrhythmic drugs- C01B	19	10	10	20	2	13	24
Anticholinergic agents	20	3	6	10	0	0	2
Beta blocker agents- C07	54	14	6	11	23	19	12
Calcium channel blockers (C08)	3	7	3	11	3	4	5
Cardiac stimulants (positive chronotropic, ionotropic, and dromotropic agents) (C01C)	7	2	6	0	0	4	11
Corticoids (S02B)	0	0	3	0	0	1	0
Digitalis glycosides (C01AA)	4	1	2	2	2	2	3
Diuretics (C03)	0	0	5	0	0	0	5
Dual anti-Platelet	72	33	29	7	45	58	38
Full anticoagulation	92	55	14	13	30	48	63
GP IIb/IIIa inhibitors	2	19	8	2	5	17	8
Lipid modifying agents (C10)	31	13	5	0	18	15	4
Magnesium	0	6	2	3	3	2	6
Miscellaneous	2	3	4	3	2	5	5
Oxygen (All other therapeutical products, V03A)	20	3	2	4	1	4	16
Plant alkaloids and other natural products (L01C)	0	1	2	0	2	1	0
Platelet aggregation inhibitors (excluding heparin) (B01AC)—single antiplatelet	88	21	20	5	27	49	43
Prophylactics and symptomatic	11	10	2	0	6	9	8
Psychotropic or anxiolytics	1	6	3	0	0	2	6
Sex hormones and modulators of the genital system (G03)/ Hormone replacement therapy	0	0	2	0	1	1	0
Thrombolytic or fibrinolytic therapy	34	21	8	12	18	21	7
Triple antithrombotic therapy	2	8	2	5	1	4	12
Vaccines (J07)	3	0	0	0	1	2	0
Vasodilators used in cardiac diseases (Nitrates)	17	3	5	6	3	5	11

Only gold members can continue reading. Log In or Register to continue

Share this:

• X
• Facebook

Related

Related posts:

1. Design and rationale of the HD PCI trial: A cluster randomized crossover trial of higher vs. lower dose heparin for elective percutaneous coronary intervention
2. Large-bore mechanical thrombectomy vs standard of care for acute high-risk pulmonary embolism: Rationale and design of the PERSEVERE randomized controlled trial
3. Comment on: Yoga vs regular exercise for atrial fibrillation: Design considerations for the yoga-AF randomized trial
4. Intravascular ultrasound guidance for complex high-risk indicated procedures: The IVUS CHIP trial study design

Stay updated, free articles. Join our Telegram channel

Join