This study is a propensity score-matching case-control type. Patients with the following criteria were included: (1) de novo TCALs defined as $\ge$25% diameter difference between the proximal and distal segments, (2) stable angina and non-ST-elevation acute coronary syndrome, and (3) available 12-month angiography. Conversely, patients with the following criteria was excluded: (1) lesions unsuitable for PCI, such as multiple-vessel disease with $>$32 syntax score, (2) ST-elevation myocardial infarction (MI) within one month, (3) patients without clinical and angiographic follow-up dada, (4) severe renal insufficiency (eGFR 30 mL/min), (5) hematopoietic disorders (platelet count 100 $\times$ 10${}^{\mathrm{^}9}$/L or $>$700 $\times$ 10${}^{\mathrm{^}9}$/L, leukocyte count 3 $\times$ 10${}^{\mathrm{^}9}$/L), (6) Intolerance to long-term antiplatelet therapy; and (7) life expectancy 1 year.

From January 2015 to May 2019, among 5055 patients who had underwent PCI, the patients who met the above criteria were matched based on propensity score matching, resulting in 150 pairs of patients treated either by MSS or COS.

All patients received pretreatments of aspirin and clopidogrel or ticagrelor with loading doses as indicated. Aspirin was maintained indefinitely, whereas clopidogrel or ticagrelor was maintained for 12 months unless contraindicated. Intra-procedural heparin of 70–100 U/kg was administered intravenously with an additional bolus of 1000 U given per hour to maintain an activated clotting time of 250–300 s. The use of platelet glycoprotein receptor antagonists was left to the discretion of the operators.

Second-generation drug-eluting stents (DESs) including Resolute (Medtronic, Minneapolis, Minnesota), Xience (Abbott Vascular, Santa Clara, CA, USA), Firebird-2 (Microport, Shanghai, China) and Excel (JW, Shandong, China) were used.

Clinical follow-up was performed through clinic visits or telephone contact at 1, 6, and 12 months after discharge and annually thereafter. Coronary angiography was planned at 12 months or performed earlier as clinically indicated. Quantitative coronary analysis was conducted in the stented segment (in-stent) and 5 mm proximal or distal to the stent end (in-edge). Restenosis was defined as $>$50% stenosis-diameter percentage at follow-up.

The primary endpoint was as follows: (1) immediate angiographic success, defined as no residual diameter stenosis $\ge$20%, abnormal TIMI flow, edge dissection $\ge$type-C, or bailout stenting; (2) major cardiac adverse events (MACEs) at one year, including cardiac death, target-vessel MI (TVMI), target lesion revascularization (TLR)/target vessel revascularization (TVR), or stent thrombosis (ST). The secondary endpoint was the MACE component.

MI was diagnosed according to the Forth Universal Definition of MI [24]. All MIs were considered as TVMI unless clear evidence indicated that they were caused by non-target vessels. TLR/TVR was repeat target vessel/lesion treatment either by PCI or CABG. ST was diagnosed according to the ARC definition [25].

Data were expressed as the mean $\pm$ SD for continuous or frequency (%) for discrete variables. To compare differences, Student’s t test was used for continuous variables, and Chi square or Fisher’s exact test was used for the discrete variables. Statistically significance was considered at P 0.05. Data were analyzed with IBM SPSS statistics (version 20.0, IBM Corp., Chicago, IL, USA).

Propensity score matching was used to reduce treatment bias and potential impact of confounding factors from baseline characteristics. All baseline clinical and lesion characteristics that may affect outcomes upon univariate analysis were deemed as candidate variables. All variables with P 0.20 were retained. Model reliability was evaluated using the Hosmer-Lemeshow test. Based on the nearest match algorithm, we created case-matched pairs without replacement at 1:1 ratio.

The baseline clinical and lesion’s characteristics (Tables 1,2) were comparable between the groups. As shown in Table 2, fewer stents were implanted (1.03 $\pm$ 0.16 mm or MSS versus 2.01 $\pm$ 0.11 mm for COS, P = 0.000) with shorter stent length per TCAL (32.08 $\pm$ 4.41 mm for MSS versus 34.42 $\pm$ 4.78 mm for COS, P = 0.012). Fewer non-compliant balloons (1.22 $\pm$ 0.47 mm for MSS versus 1.78 $\pm$ 0.55 mm for COS, P = 0.000) were used for post-dilation in the MSS group. Immediate angiographic success was comparable between the groups (96.7% for MSS versus 98.0% for COS, P = 0.723) with similarly low rates of edge dissection (2.7% for MSS versus 1.3% for COS, P = 0.684) and edge bailout stenting (2.7% for MSS versus 1.3% for COS, P = 0.684). Additionally, procedural time, radiation dosage, contrast volume and treatment cost per lesion were reduced in the MSS group.

Table 3 shows that the baseline lesion characteristics were similar between the groups. Immediately after the procedure, we observed smaller in-stent MLD (2.96 $\pm$ 0.34 mm for MSS versus 3.08 $\pm$ 0.33 mm for COS, P = 0.004) and higher diameter stenosis (11.7 $\pm$ 4.0% for MSS versus 9.0 $\pm$ 4.8% for COS, P = 0.003) in the MSS group, as well as similar in-edge MLD and diameter stenosis between the groups. At 1-year follow-up, we observed larger in-stent MLD (2.76 $\pm$ 0.38 mm for MSS versus 2.65 $\pm$ 0.60 mm for COS, P = 0.003), less LLL (0.20 $\pm$ 0.26 mm for MSS versus 0.42 $\pm$ 0.48 mm for COS, P = 0.001), less diameter stenosis (19.68 $\pm$ 11.75% for MSS versus 24.02 $\pm$ 20.94% for COS, P = 0.028), and fewer binary restenosis (10.0% for MSS versus 18.7% COS, P = 0.047) in the MSS group. The in-edge MLD, LLL, diameter stenosis, and binary restenosis were similar between the groups.

Angiographic success (96.7% versus 98.0%, P = 0.723) was comparable between the MSS group and COS groups. The one-year cumulative MACE (12.0% versus 22.7%, P = 0.022) and TLR/TVR (10.0% versus 18.7%, P = 0.047) were significantly reduced in the MSS group compared with the COS group. No difference in cardiac death, TVMI, and ST was observed between the groups (Table 4).

Current strategies to treat TCALs include overlapped stenting with multiple short stents [26, 27, 28], single stenting with a conventional long tubular stent [29, 30], or single stenting with a tapered long stent [18, 19, 20, 21, 22, 23].

Overlapped stenting, applied as early as the era of bare-metal stents (BMS), remains the most common treatment for TCALs. The benefit of stent overlapping is that it can match a long tapered vessel by stepping up the size of multiple stents. However, clinical outcomes afforded by overlapping BMS have been proven inferior to those treated with a single BMS primarily due to increased TLR [31, 32, 33, 34]. Overlapping stents with first generation DESs could effectively reduce restenosis by strongly inhibiting neointimal hyperplasia [35, 36, 37, 38, 39] However, clinical outcomes remain controversial. A pooled analysis of five studies on overlapping sirolimus-DESs has revealed that the rates of ischemic end-points and revascularization are similar to those of a single sirolimus-DES, and that revascularization is significantly reduced compared with a BMS [40]. By contrast, a study comparing overlapping DESs, non-overlapping DESs, and a single DES implanted in a vessel has demonstrated that overlapping DESs are associated with impaired angiographic and long-term clinical outcomes, including death or MI [41]. The discrepancy could be partly explained by the delayed vascular healing and impaired endothelialization caused by increased drug concentrations and polymer burden because impaired endothelialization is particularly pronounced at overlapped-stent sites [42]. Additionally, an experimental study has shown more neutrophils, eosinophils, and fibrin deposition at the sites of overlapping DESs than at those of non-overlapping DESs and BMSs. This finding suggests the inflammation of impaired vascular healing at DES overlapping sites [43]. Overall, these data suggest that a single long stent may be better than multiple overlapping stents for the treatment of long lesions or TCALs. Obviously, a long tapered stent may be more suitable for the fixation of long TCALs. Nevertheless, tapered or long-tapered stents are not extensively used clinically because of their limited availability.

Considering the inconsistent outcomes of overlapping stents and the limited availability of long tapered stents, most treatment of TCALs involve the use of the conventional tubular long stent. Theoretically, the expandability of conventional tubular stents has a maximum limit despite an open-cell design. Further dilation with larger balloons and higher pressure inevitably deforms the stent structure and disrupt the stent polymer, thereby leading to likely unfavorable outcomes [6, 7, 8, 9, 10, 44, 45]. Apart from the use of overlapping multiple stents or tapered stents, this dilemma remains unsolved.

In the present study, we used an MSS characterized by two key points: First was the selection of a larger stent based on the mean distal and proximal RVD instead of the distal RVD to obtain a larger expandable lumen and to more effectively adapt the tapered anatomy of TCALs without deformation the stent platform; The second key point was stepwise stent deploying by initially inflating with a low pressure of about 6 atm (much lower than the nominated pressure). This process was followed by reinflating with the nominated pressure or a higher one after pulling back the stent balloon by 1–2 mm to avoid distal edge dissection caused by deploying an oversized stent. As shown in our study, the stent selected using such a standard was larger than that using the conventional one. Our study further showed that the MSS achieved better clinical outcomes than OSMS, as evidenced by the lower rate of MACE at 1 year, less use of stents, and lower cost of treatment with similar rates of immediate angiographic success.