The Assessment of Indications for Percutaneous Coronary Intervention

Summary

Background:

Percutaneous coronary intervention (PCI) is performed 300 000 times a year in Germany. Half of these procedures involve patients with chronic coronary syndrome. Little information is available on the quality of decision-making about the indication for PCI in this patient group.

Methods:

The Center for Health Services Research of the German Cardiac Society (Zentrum für kardiologische Versorgungsforschung der Deutschen Gesellschaft für Kardiologie – Herzund Kreislaufforschung e.V., DGK-ZfKVF), in collaboration with the Initiative on Quality in Medicine (Initiative Qualitätsmedizin e.V., IQM) carried out this study, in which the indication for PCI in randomly selected patients with chronic coronary syndrome was reviewed by peers on site.

Results:

441 assessment forms from 16 hospitals were included in the study. More than half of the patients (52.2%) had typical chest pain, 20.1% had mainly dyspnea, 17.7% had chest pain and dyspnea, and 10% had no symptoms. Three-quarters had symptoms or ischemia despite optimal drug treatment. Most patients were treated for clinical symptoms and proven ischemia or stenosis of more than 90%. In 11% of patients, the indication was complete revascularization as a staged procedure after a prior myocardial infarction treated with PCI. 73.7% of indications for PCI were judged to be in adherence to the guidelines; guideline adherence was judged to be questionable in 24.1% and to be absent in 2.2%.

Conclusion:

These data on the quality of indications for elective PCI reveal a need for improvement in adherence to the recommendations of the current German and international guidelines. In particular, the non-invasive or invasive proof of ischemia should be performed more frequently. These findings underscore the importance of peer review of indications for interventional procedures.

Much effort is expended on striving to describe and compare the quality of medical service provision (1). While this is relatively simple for structures and processes, when it comes to assessing outcome quality one crucial dimension is usually missing: the quality of decisionmaking with regard to indications. However, indication quality is indubitably an indispensable factor in the evaluation of performance and the comparison of service providers. Documentation of indication quality is complex and cannot be achieved by use of routinely recorded parameters: elaborate on-site audits or peer reviews are required (2, 3). Percutaneous coronary intervention (PCI) is one of the most commonly performed medical procedures, with over 300 000 cases annually in Germany (4). Around half of these interventions are carried out in patients with chronic coronary syndrome (5). In terms of interventions per million inhabitants, PCI is performed more frequently in Germany than in any other European country (6). In Sweden, for example, the rate of PCI is about 40% lower.

While there is strong evidence for the symptomatic and prognostic benefit of PCI in acute coronary syndrome (7), in patients with chronic disease PCI predominantly ameliorates the symptoms (8). The indications for PCI in chronic coronary syndrome are defined in the guidelines of the European Society of Cardiology (ESC) (9, 10) and in the German National Disease Management Guideline (Nationale Versorgungsleitlinie, NVL) on chronic coronary heart disease (11). The aim of the study presented here was to evaluate adherence to these guidelines in German clinical practice.

Methods

This study was a collaboration between the Center for Health Services Research, German Society of Cardiology – Cardiovascular Research (DGK-ZfKVF), the Initiative for Quality in Medicine (IQM), and the Institute for Myocardial Infarction Research Ludwigshafen. It was not entered prospectively in a study registry.

Sixteen 16 IQM peer reviews were conducted by IQM peers in randomly selected IQM-affiliated hospitals. The peer reviews were carried out in accordance with the standard IQM methods. Members of the IQM are obliged to apply this methodology , which is based on quality assessment, transparency, and peer review. In the course of the peer review process, certain defined circumstances, such as a change of medical director or departmental restructuring, may provide grounds for an appeal against the review. In the period 2022–2024, 11 of 27 planned peer reviews were canceled for such reasons. According to the DGK statistics for the year 2022, 696 German hospitals performed PCI in a total of 178 785 cases (12). In 2021, 310 of these hospitals affiliated to the IQM carried out at least one PCI, with 74 012 cases altogether. IQM hospitals therefore represent 44.5% of the hospitals that offer PCI in Germany, while 41.4% of PCIs in Germany are performed in IQM-affiliated hospitals.

The peer reviews were conducted exclusively in hospitals where at least 30 PCIs had been carried out in 2021. This inclusion criterion was met by 247 institutions (35.5%) with a total of 73 687 cases of PCI (41.2%). All IQM-affiliated hospitals provide physicians for the peer review process, all of whom are qualified as IQM peers by virtue of successful completion of the German Medical Association’s “Peer Review in Medicine” curriculum. All IQM peer physicians are heads of department, medical directors, or senior physicians. The peer review teams generally consisted of four to five peers: a team leader with anesthesiological expertise and and three to four peers with expertise in interventional cardiology. In addition to specialist qualification, the factors considered in putting together the teams were geographical distance from the hospital being visited, the situation regarding competition, and the peers’ degree of experience with the peer review process. The peer review teams were assigned centrally by the Peer Review Steering Committee of the IQM. At each hospital, the cases of up to 30 randomly selected patients with PCI and chronic coronary syndrome were assessed by the IQM peers in on-site peer reviews. The IQM peers also evaluated the images obtained in the course of coronary angiography and PCI. Furthermore, cases coded as coronary artery dilation and stent implantation (8 837 ff.) in the German classification of operations and procedures (OPS) were selected, excluding patients with acute myocardial infarction (ICD 25.1–4). The participating hospitals gave the peers access to the case records, and 30 anonymized patient files were selected for peer review. In all IQM peer reviews, case selection took place centrally: the IQM cooperation partner and data trustee Solventum Germany GmbH selected cases according to criteria defined by the Peer Review Steering Committee. All selected records, including the respective hospital’s own assessments, were provided to the IQM peers. Cases were selected for peer review using the denominator of the German Inpatient Quality Indicator (G-IQI) 3 132, “Therapeutic catheterization of coronary vessels without main diagnosis of myocardial infarction, without a surgical procedure from sector 5 of the OPS

(age > 19), fraction fatalities”. The G-IQIs are a diseasespecific system of indicators based on routine hospital claims data and used to portray the quality of inpatient care in all IQM-affiliated hospitals. Initially, pseudonymized case numbers with no relation to the patients were selected on the basis of the denominator of this indicator, excluding cases that did not meet the additional inclusion criterion “simultaneous presence of chronic coronary heart disease (CHD).” The physicians who had determined the indications were interviewed at the conclusion of the peer review; however, their responses were not included in our analyses.

A guideline-adherent medical therapy (9, 11) before PCI was defined and comprised two prognostically and symptomatically beneficial (antithrombotic, LDL-lowering) drugs and at least one antianginal substance.

The indication for PCI was assessed according to the criteria of the current ESC guidelines on revascularization and the NVL on chronic CHD (Table 1). While the ESC guidelines make recommendations regarding stenosis severity and testing for ischemia, the NVL states: “Revascularization should be offered in the presence of symptoms not sufficiently controlled by conservative treatment, provided invasive diagnostic work-up shows the patient’s coronary anatomy is suited for PCI.” (11). Differences between the two sets of recommendations refer essentially to the emphasis on optimal drug therapy before PCI in the NVL and the emphasis on stenosis severity and ischemia testing in the ESC guidelines. The overall assessment took both guidelines into account (Table 1).

Table 1. Definitive and possible indications for percutaneous coronary intervention in persons with chronic coronary syndrome^*.

Definitive indication	Possible indication
Symptomatic patient (chest pain or dyspnea)
•Positive non-invasive ischemia test and/or •Positive pressure wire measurement and/or •≥ 90% stenosis and/or •Persistence of symptoms despite optimal medical therapy plus anatomy compatible with PCI	≥ 70% stenosis
Asymptomatic patient
•Positive non-invasive ischemia test and/or •Positive pressure wire measurement	≥ 90% stenosis
Complete revascularization within 45 days after PCI in myocardial infarction
≥ 70% stenosis	≥ 50% stenosis

^*Modified from the NVL on chronic CHD (11) and the European Society of Cardiology’s guidelines (9)

PCI, Percutaneous coronary intervention

If two definitive criteria (e.g., symptoms and positive ischemia testing) were met, PCI was classed as definitively indicated. Fulfillment of one primary and one possible criterion led to classification of the indication for PCI as possible, and if no primary or possible criteria were met, then the intervention was deemed not indicated. In addition, complete revascularization on two separate occasions after PCI in a patient with acute myocardial infarction was classed as a guideline-adherent indication (7).

The Institute for Myocardial Infarction Research Ludwigshafen had access to a completely anonymized version of the database (neither hospital nor patient/case could be identified) for evaluation.

Statistics: the planned sample size of > 450 cases was adequate for estimation of cases with non-guidelineadherent indication with accuracy of ± 3.2%—assuming that the proportion of such cases would be in the region of 10%.

The predominantly bivalent (yes/no) and categorical variables were summarized by means of absolute frequencies and percentages and tested for differences between the indication groups using Pearson’s chi-squared test. Continuous variables—age, serum creatinine, estimated glomerular filtration rate (eGFR)—were expressed as mean with standard deviation and 95% confidence interval, and the group differences were tested using Wilcoxon’s rank sum test. All statistics are based exclusively on the data provided, with no imputation. The statistical software used was SAS Version 9.4 for Windows (© SAS Institute Inc.).

Results

Between October 2023 and December 2024, a total of 474 randomly selected cases (28–30 from each of 16 hospitals) from 2023 and 2024 were reviewed on site by IQM peers and evaluated using a standardized questionnaire. Thirty-three cases had to be excluded from analysis (acute myocardial infarction n = 17; no PCI n = 4; assessment impossible due to insufficient documentation of clinical symptoms and/or ischemia diagnosis n = 12).

The basic variables of the 441 patients analyzed are shown in Table 2. More than half of them (52.2%) had typical chest pain, while 20.1% suffered predominantly from dyspnea, 17.7% displayed both chest pain and dyspnea, and 10% had no symptoms. The pharmacological treatment before PCI comprised antithrombotic substances in 90% of cases, lipid-lowering drugs in 85%, and anti-anginal agents in 89.7% (Table 2). Thus 75.9% of the patients were medicated in accordance with the guidelines, receiving drugs of three different types prior to PCI. The documented testing for ischemia (Table 3) comprised ergometry in 10.4% of those treated, regional ischemia testing in 35.1%, and invasive pressure wire measurement in 16.8%. The severity of the treated coronary stenosis before PCI was ≥ 90% in 50.3% of the patients, 70–89% in 37.2% of cases, and < 70% in 12.5%.

Table 2. The basic variables of the 441 patients with chronic coronary syndrome.

	Patients (n)	% [95% CI]
Age in years, mean ± SD	69.8 + 13.5
Number of women	115/423	27.2 [23.0; 31.7]
Diabetes mellitus	151/407	37.1 [32.4; 42.0]
Previous PCI	227/406	55.9 [50.9; 60.8]
Status post myocardial infarction	142/394	36.0 [31.3; 41.0]
Left ventricular ejection fraction
–normal (> 50%)	225/375	60.0 [54.8; 65.0]
–slightly reduced (40–50%)	72/375	19.2 [15.3; 23.6]
–severely reduced (< 40%)	78/375	20.8 [15.1; 27.8]
Status post stroke	34/392	8.7 [6.1; 11.9]
Renal insufficiency with eGFR < 60 mL/min/1.73 m2	132/380	34.7 [30.0; 39.8]
Chest pain	354/441	80.3 [76.2; 83.9]
No symptoms	44/441	10.0 [7.3; 13.2]
Drug treatment according to the guidelines	306/403	75.9 [71.5; 80.0]
Persistent symptoms despite treatment according to the guidelines	170/304	55.9 [50.1; 61.6]

CI, Confidence interval; eGFR, estimated glomerular filtration rate; n, number; PCI, percutaneous coronary intervention; SD, standard deviation

Table 3. Non-invasive and invasive testing for ischemia.

Method	n	% [95% CI]
Positive ergometry	46	10.4 [7.7; 13.7]
Localized testing	155	35.1 [30.7; 39.8]
Invasive pressure wire measurement performed	118	26.8 [22.7; 31.1]
Invasive pressure wire measurement positive	36	8.2 [5.8; 11.1]
Intracoronary imaging (IVUS/OCT)	28	6.3 [4.3; 9.0]

CI, Confidence interval; IVUS, intravascular ultrasonography; n, number; OCT, optical coherence tomography

The indications for PCI in the whole group are shown in the Figure. Most patients (75.9%) who underwent PCI had clinical symptoms or ischemia despite optimal medical therapy. Table 4 shows our analysis of guidelineadherent indications for PCI for the different combinations of findings. The overwhelming majority of patients treated with PCI had clinical symptoms with demonstrated ischemia or with > 90% stenosis. PCI in asymptomatic persons conformed with the guidelines in only 25 of 41 cases.

Figure.

Indications for percutaneous coronary intervention in the 441 patients with chronic coronary syndrome (multiple entries possible) OMT, Optimal medical therapy

Table 4. Guideline-adherent indications for percutaneous coronary intervention after assessment by peer review (n = 325/441; 73.7%).

Criteria	n	% [95% CI]
Complete revascularization after myocardial infarction	43/441	9.3 [6.5; 12.7]
Chest pain or dyspnea + positive test for ischemia	65/441	14.7 [10.8; 17.9]
Chest pain or dyspnea + positive pressure wire measurement	20/441	4.5 [2.8; 6.5]
Chest pain or dyspnea + ≥ 90% stenosis	128/441	29.0 [25.0; 33.7]
Chest pain or dyspnea + 70–90% stenosis	47/441	10.7 [7.1; 12.8]
Chest pain or dyspnea despite optimal medical therapy + > 70% stenosis	11/441	2.5 [1.2; 3.8]
No symptoms + positive test for ischemia	11/441	2.5 [1.2; 3.8]
No symptoms + positive pressure wire measurement	2/441	0.5 [0.1; 1.2]
No symptoms + ≥ 90% stenosis	11/441	2.5 [1.2; 3.8]

CI, Confidence interval; n, number

According to the peers’ assessment, the indication for PCI adhered to the guidelines in only 73.7% of the patients. In 24.1% the indication was classed as questionable, and in 2.2% of cases the peers were of the opinion that PCI was not indicated at all. The center where PCI took place had no effect on whether the indication was guideline-adherent. The indications conformed with the guidelines in 75.7% of women and 72.4% of men (p = 0.50). Considering the ESC guideline alone, 58.5% of the indications were classed as guideline-adherent and 37.4% as questionable. Taking into account only the criteria of the NVL on chronic CHD, 53.7% of the indications conformed with the guidelines and 43,5% were questionable. There are no data on inter-rater agreement, because the peers were also anonymized.

Discussion

The guidelines on the indication for PCI contain information on the following:

• The necessary clinical symptoms

• The recommended anti-anginal and prognostic medication before PCI

• The non-invasive and invasive tests for ischemia

• The severity of stenosis before PCI (9–11)

Ideally, to be a candidate for PCI a patient should have chest pain or dyspnea that cannot be managed with medication (11). Moreover, the presence of ischemia should be demonstrated by ergometry—or better still, regionally by means of stress echocardiography, myocardial scintigraphy, or stress MRI. Alternatively, the hemodynamic significance of the ischemia can be assessed invasively by means of pressure wire testing, e.g., measurement of the fractional flow reserve (FFR), the instantaneous wave-free ratio (iFR), or the resting full-cycle ratio (rFR). The number of PCIs per million inhabitants is higher in Germany than in other European countries (6). Whether this means there is overuse of PCI throughout Germany or underuse elsewhere is currently a topic of debate (13). This question can best be answered by examining to what extent decisions on indication follow the guidelines.

The documentation of indication quality is a crucial precondition for the assessment of medical service provision. While there have been numerous attempts to document and evaluate the quality of conduct and outcome of services provided in Germany (1), the assessment of indication quality is much less frequent. One reason for this is that reliable quality assessment of the indications for medical interventions can only be achieved by time-consuming and staff-intensive on-site audits or peer reviews (2, 3). PCI was one of the first medical procedures for which the indications were assessed in this way, in a pilot project of the ALKK (Arbeitsgemeinschaft Leitende Kardiologische Krankenhausärzte) (14). The intention was primarily to document outcome quality (15), but indication quality was also evaluated. A central commission analyzed 381 cases of PCI and concluded that the quality of the indication for PCI was:

• Indubitable in 40% of cases

• Adequate in 41% of cases

• Questionable in 15% of cases

• Inadequate in 4% of cases (2)

These findings are comparable with the results of the study reported here. Thus the proportion of nondefinitive indications for PCI in chronic coronary syndrome seems not to have changed greatly over the past 20 years.

A large analysis of over 500 000 PCIs in the USA included 144 737 interventions in patients with chronic coronary syndrome (16). In this study 50.4% of the PCIs were classified as clearly indicated, 38.0% as questionable, and 11.6% as not indicated. The rate of non-definitively indicated PCIs was higher in an American analysis of 33 970 such interventions in patients without acute coronary syndrome (17). Here, 14.3% of PCIs were not indicated, while the indication for as many as 49.6% was questionable. A metaanalysis published in 2023 revealed that PCI was definitively not indicated in 4.3% of cases in which it was performed in patients with acute coronary syndrome, while the rate in those with chronic coronary syndrome was 8.9%. These findings are much higher than our observed rate of 2.3% of definitively non-indicated PCIs. However, neither of these two studies classified the PCIs by means of on-site audits or peer reviews.

Complete and readily comprehensible documentation of the clinical symptoms and the invasive and non-invasive diagnostic procedures is crucial. Only then can the quality of decisions on indication be assessed. In our study, 12 of 474 cases (2.5%) had to be excluded because the available records did not permit such assessment. Moreover, it is possible that the 24.2% of questionable indications include cases in which, for example, prior diagnostic testing for ischemia was not properly documented in the hospital files. It may also be that bottlenecks in timely availability of non-invasive coronary diagnostics outside the hospital setting—particularly in rural areas—may have contributed to the tendency for patients with typical symptoms to undergo direct coronary diagnostic procedures. In accordance with the recommendation of the NVL on chronic CHD (11), in the analysis presented here PCI was most often carried out in persons with persisting clinical symptoms despite optimal medical therapy. This documents good adherence to this German national guideline.

There were no differences among centers in decisions on indications, nor did women and men differ in the guideline-adherent indication for PCI.

A limitation of our study is the small sample size compared with the total number of PCIs. Selection bias on the part of the hospitals cannot be excluded, since they are members of the Initiative for Quality in Medicine. On the other hand, the IQM-affiliated hospitals perform around 40% of all PCIs in Germany.

The findings reported here demonstrate the necessity of peer review for adequate assessment of the quality of the decisions on indications. Improvement was found to be needed both in the conduct and complete documentation of PCI and in the complete documentation of the preceding non-invasive procedures and the clinical symptoms. Fewer than 3% of the interventions were definitively not indicated, while around three fourths were definitively indicated. The goal should be to reduce the size (currently around 24%) of the gray area in between. The best way of achieving this aim would be continuous monitoring of indication quality, e.g., by means of peer reviews carried out by the IQM. It would be desirable for this work to be undertaken by medical professionals, to ensure the greatest possible competence in this key area of quality assurance.