Abstract
Introduction:
The coronavirus disease 2019 (COVID-19) pandemic’s impact on vascular procedural volumes and outcomes has not been fully characterized.
Methods:
Volume and outcome data before (1/2019 – 2/2020), during (3/2020 – 4/2020), and following (5/2020 – 6/2020) the initial pandemic surge were obtained from the Vascular Quality Initiative (VQI). Volume changes were determined using interrupted Poisson time series regression. Adjusted mortality was estimated using multivariable logistic regression.
Results:
The final cohort comprised 57,181 patients from 147 US and Canadian sites. Overall procedure volumes fell 35.2% (95% CI 31.9%, 38.4%, p < 0.001) during and 19.8% (95% CI 16.8%, 22.9%, p < 0.001) following the surge, compared with presurge months. Procedure volumes fell 71.1% for claudication (95% CI 55.6%, 86.4%, p < 0.001) and 15.9% for chronic limb-threatening ischemia (CLTI) (95% CI 11.9%, 19.8%, p < 0.001) but remained unchanged for acute limb ischemia (ALI) when comparing surge to presurge months. Adjusted mortality was significantly higher among those with claudication (0.5% vs 0.1%; OR 4.38 [95% CI 1.42, 13.5], p = 0.01) and ALI (6.4% vs 4.4%; OR 2.63 [95% CI 1.39, 4.98], p = 0.003) when comparing postsurge with presurge periods.
Conclusion:
The first North American COVID-19 pandemic surge was associated with a significant and sustained decline in both elective and nonelective lower-extremity vascular procedural volumes. When compared with presurge patients, in-hospital mortality increased for those with claudication and ALI following the surge.
Introduction
The coronavirus disease 2019 (COVID-19) pandemic, caused by the severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2), arrived in the United States and Canada in early 2020,1,2 and resulted in severe disruption of their respective health care systems. Elective cardiovascular procedures were temporarily halted3,4 at many centers and unintended reductions in urgent and emergent presentations were observed as well. 5 Although declines in lower-extremity revascularization volumes have been reported in surveys of US vascular specialists,6–8 the full impact of COVID-19 on vascular procedural volumes and outcomes in North America during the pandemic surge has yet to be characterized.9,10 We sought to assess endovascular and open surgical lower-extremity revascularization and amputation procedural volumes and outcomes before, during, and after the first North American COVID-19 pandemic surge, using data from the Society for Vascular Surgery Vascular Quality Initiative (VQI). 1
Materials and methods
Data source
The VQI is a distributed network of quality groups functioning under the Society for Vascular Surgery Patient Safety Organization (SVS-PSO). Its 14 registries include data from > 700 participating centers, > 700,000 procedures from academic (58%) and community (42%) medical centers, performed by > 3500 multispecialty physicians including vascular surgeons (46%), interventional cardiologists (15%), interventional radiologists (15%), general surgeons (5%), and cardiothoracic surgeons (5%). 11 The present study employed data from the Peripheral Vascular Intervention, Supra-Inguinal Bypass, Infra-Inguinal Bypass, and Amputation registries. Each of these modules prospectively collects > 100 demographic, comorbid, procedure, and outcome variables.12–14
Patients
Patients aged 18 years or older who underwent supra-inguinal bypass, infra-inguinal bypass, endovascular peripheral vascular intervention (PVI) or amputation between January 1, 2019 and June 30, 2020 were included. Sites with a pre-COVID-19 volume of < 10 cases/month and those sites not submitting data for each consecutive month during the pandemic were excluded. The Rhode Island Hospital Institutional Review Board determined that this study did not constitute human subjects research, and the requirement for informed consent was therefore waived. Subject COVID-19 status (i.e., positive, negative, suspected) was not available.
Exposure
COVID-19 was officially declared a global pandemic by the World Health Organization (WHO) on March 11, 2020. Monthly procedure volumes during the study period for each of the 18 participating North American VQI regions appear in online Supplemental Figure 1. Presurge, surge, and postsurge periods were defined as January 1, 2019 – February 29, 2020, March 1 – April 30, 2020, and May 1 – June 30, 2020, respectively.
Outcomes
The primary study outcome was procedural volume. Monthly volumes were determined in the overall cohort and by registry module and categorized as elective versus nonelective. Relative volume per center per month during the surge was compared with that during presurge and postsurge periods in the overall cohort, and separately by registry, urgency, presenting symptoms, and geographic region. Finally, differences between observed and expected volumes during surge and postsurge months were determined for the overall cohort and separately by presenting symptom. Secondary outcomes included in-hospital death, composite death or myocardial infarction (MI), pulmonary complications, renal complications, postprocedure amputation, total length of stay (LOS), and discharge disposition. Select definitions of postoperative complications are shown in online Supplemental Table 3.
Statistical analysis
Continuous variables are displayed as means with SD and compared using t-tests, or as medians with IQR and compared with Mann–Whitney tests if non-normally distributed. Categorical variables are displayed as frequencies/percentages and compared using chi-squared or Fisher’s exact test as appropriate. Interrupted Poisson time series regression was used for comparisons of procedure volume per center per month. 15 A sensitivity analysis of procedure volume over time was performed after restricting the analytic cohort to those whose records were entered in the registry within 60 days of the procedure date. The rationale for doing so was as follows: (1) records can be created at any time following procedures in VQI modules; and (2) the opportunity to create records for procedures performed at the end of the study period was no greater than 60 days. A Poisson regression model was used to compute expected counts during surge and postsurge months (i.e., March, April, May, and June 2020) based on volumes from presurge months. Differences between observed and expected monthly procedure volumes were computed and displayed graphically. A p-value less than 0.05 was considered statistically significant. SAS software, Version 9.4 (SAS Institute Inc., Cary, NC, USA) was used for all analyses.
Results
Patient characteristics
There were 76,542 patients who underwent lower-extremity vascular procedures or amputations at 326 sites during the study period. After excluding sites whose nonsurge volume was < 10 cases/month (n = 147 sites; n = 9609 patients) and those not submitting data during each month of the study period (n = 32 sites; n = 9752 patients), 57,181 patients from 147 sites remained (Figure 1). The final cohort was comprised of 43,538 endovascular PVI, 2301 supra-inguinal bypass, 7237 infra-inguinal bypass, and 4105 amputation patients. Of these, 46,960 (82.1%) underwent procedures presurge, 4719 (8.3%) during the surge, and 5502 (9.6%) following the surge. Patient demographic and comorbid and treatment characteristics before, during, and following the COVID surge for the overall cohort and separately by procedure registry appear in Table 1. Patient age and sex were similar during presurge, surge, and postsurge periods; however, patients presenting during the surge were less often White, more likely to have diabetes, heart failure or undergo dialysis, and more likely to present with chronic limb-threatening ischemia (CLTI) or acute limb ischemia (ALI).
Study flow diagram.
Patient and procedure characteristics before, during, and after the COVID-19 surge.
COVID-19, coronavirus disease 2019; AKA, above-knee amputation; AK pop, above-knee popliteal; ALI, acute limb ischemia; ASA, American Society of Anesthesia; AT, anterior tibial; BKA, below-knee amputation; BK pop, below-knee popliteal; CHF, congestive heart failure; CLTI, chronic limb-threatening ischemia; Com fem, common femoral; COPD, chronic obstructive pulmonary disease; DP, dorsalis pedis; Ext, external; PT, posterior tibial; PVI, peripheral vascular intervention; SFA, superficial femoral artery; TASC, Trans-Atlantic Inter-Society Consensus; TKA, thru-knee amputation; TMA, trans-metatarsal amputation; TP, tibioperoneal.
Procedure characteristics
Procedural characteristics before, during, and following the surge for the overall cohort, and separately by procedure registry, appear in Table 1.
PVI
Among patients undergoing PVI, fluoroscopy times and contrast volumes were similar during the surge, compared with pre- and postsurge periods. However, patients undergoing PVI during the surge were more likely to undergo procedures for CLTI and ALI, and less likely to undergo these for claudication compared with before or after the surge. PVI procedures were performed less often for inflow and more often for infra-popliteal disease, involved a greater number of vessels, were of greater Transatlantic Inter-Society Consensus (TASC) complexity, longer, more calcified, and required a greater number of devices when comparing surge months with those before or after the surge.
Surgery
For supra- or infra-inguinal bypass, procedures were more often performed for CLTI and ALI than claudication during the surge compared with pre- and postsurge periods. Supra-inguinal bypass procedures were longer during and following than before the surge; there was no difference in the need for concomitant PVI across the three periods. Among those undergoing infra-inguinal bypass, the use of general anesthesia and the incidence of concomitant PVI increased during surge and postsurge compared with presurge months. Amputation was performed less often for manifestations of CLTI (specifically, ischemic rest pain and tissue loss) and more often for ALI, uncontrolled infection, and neuropathic tissue loss during and following than before the surge. The level of amputation also varied over time such that below-the-knee amputations were less common and above-the-knee amputations more common during than before or after the surge.
Procedure volumes
Monthly procedure volumes for the overall cohort, and separately for patients undergoing PVI, supra-inguinal bypass, infra-inguinal bypass, and amputation appear in Figure 2; a steep decline in volumes occurred during surge months, which was followed by a rebound immediately thereafter. The proportion of nonelective cases rose sharply during the surge and began returning toward baseline thereafter. The volume during and after the surge was compared with that before the surge for the overall cohort, and separately by registry, urgency, presenting symptom, and geographic region appear in Table 2. In the overall cohort, volumes fell by 35.2% (95% CI 31.9%, 38.4%, p < 0.001) during the surge and by 19.8% (95% CI 16.8%, 22.9%, p < 0.001) postsurge, compared with presurge months. When comparing surge with presurge months, volume reductions from greatest to smallest were observed for supra-inguinal bypass (–46.6% [95% CI –63.7%, –29.6%], p < 0.001), PVI (–37.3% [95% CI –41.1%, –33.6%], p < 0.001), infra-inguinal bypass (–29.5% [95% CI –38.5%, –20.6%], p < 0.001), and amputation (–17.8% [95% CI –29.2%, –6.5%], p = 0.005), respectively. Procedures performed for claudication fell by 71.1% (95% CI –86.4%, –55.6%, p < 0.001), while those for CLTI fell by 15.9% (95% CI –19.8%, –11.9%, p < 0.001); those for ALI remained unchanged when comparing surge to presurge months. US geographic regions were variably affected; when comparing surge with presurge months, the largest to smallest volume reductions were observed in the East (–41.5% [95% CI –46.9%, –36.1%], p < 0.001), North (–36.3% [95% CI –43.3%, –29.3%], p < 0.001), South (–29.5% [95% CI –37.7%, –21.4%], p < 0.001) and West (–14.5% [95% CI –25.0%, –4.1%], p = 0.01), respectively. In a sensitivity analysis of overall volumes, where the analytic cohort was restricted to records entered ⩽ 60 days after the procedure, volumes during the surge were reduced by 29.9% (95% CI 26.3%, 33.5%, p < 0.001) when compared with presurge months; this reduction was not significantly different than that observed in the overall cohort. In contrast, procedure volumes were reduced by only 5.2% (95% CI 1.9%, 8.4%, p = 0.004) immediately following the surge compared with presurge months; this reduction was significantly smaller than that observed in the overall cohort. Finally, observed versus expected monthly procedure volumes for the overall cohort and by presenting symptom are depicted in Figure 3. Reductions in anticipated procedures were greatest during April and May 2020 and occurred more so for claudication than CLTI; observed and expected procedure volumes for ALI were similar over time. There were 671 ‘lost’ (i.e., expected minus observed) procedures for CLTI during these 2 months (online Supplemental Table 2).
Monthly procedure volume by registry and % nonelective in overall cohort.
Procedure volume during COVID-19 surge and postsurge versus presurge periods (n = 57,181).
COVID-19, coronavirus disease 2019; ALI, acute limb ischemia; Amp, amputation; CLTI, chronic limb-threatening ischemia; Infra, infra-inguinal bypass; PVI, peripheral vascular intervention; Supra, supra-inguinal bypass.
Unadjusted in-hospital outcomes in the overall cohort and by presenting symptom before, during, and after COVID-19 surge.
COVID-19, coronavirus disease 2019; ALI, acute limb ischemia; CLTI, chronic limb-threatening ischemia; LOS, length of stay; MI, myocardial infarction.
Observed minus expected procedure volumes by study month.
Secondary outcomes
Unadjusted in-hospital complications for the overall cohort and according to presenting symptom appear in Table 3. Median (IQR) LOS was significantly greater during (0 [0–5] days) and following (0 [0–5] days) compared with before (0 [0–4] days) the surge, although these differences were small. Discharge disposition shifted away from nursing homes toward home during the surge and shifted again from home toward rehabilitation facilities following the surge. In the overall cohort, unadjusted mortality rates rose significantly during and following the surge compared with presurge months. In a mortality model adjusting for patient characteristics, an interaction was observed between presenting symptom and time period (p = 0.031). Adjusted mortality, stratified by presenting symptom, was significantly higher during postsurge compared with presurge months for those with claudication (0.5% vs 0.1%; OR 4.38 [95% CI 1.42, 13.5], p = 0.01) and ALI (6.4% vs 4.4%; OR 2.63 [95% CI 1.39, 4.98], p = 0.003); no difference in mortality was observed for those presenting with CLTI (Table 4). Unadjusted registry-specific outcomes appear in online Supplemental Table 1. Post-PVI amputation and planned amputation were more common during surge than nonsurge months, but the proportion of amputations that were planned remained unchanged. In patients undergoing PVI, thrombotic complications occurred more often during than before or after the surge (1.1% vs 0.8% and 0.8%, respectively, p = 0.03). Among those undergoing supra-inguinal bypass, stroke occurred more commonly during and after than before the surge (1.2% and 3.2% vs 0.7%, respectively, p = 0.001). Patients undergoing infra-inguinal bypass were more likely to develop heart failure during admission following than before or during the surge (2.3% vs 1.2% and 1%, respectively, p = 0.025). No differences in complication rates were seen over time in patients undergoing amputation. Changes in LOS varied across procedural registries, with those for PVI increasing and those for amputation decreasing during surge and postsurge compared with presurge months.
Adjusted in-hospital mortality in the overall cohort and by presenting symptom before, during, and after COVID-19 surge.
Other model covariates (OR [95% CI]): age, per 10-year increase (1.27 [1.14, 1.4], p < 0.0001); male sex (0.75 [0.6, 0.94], p = 0.012); Hispanic ethnicity (1.31 [0.87, 2], p = 0.2); prior vs never smoker (0.94 [0.71, 1.24], p = 0.67); current vs never smoker (0.99 [0.72, 1.35], p = 0.95); dialysis (2.66 [2, 3.53], p < 0.001); prior congestive heart failure (1.56 [1.22, 1.99], p = 0.0004); symptom*time interaction: p = 0.031.
Presurge, surge, and postsurge periods were defined as January 1, 2019 – February 29, 2020, March 1 – April 30, 2020, and May 1 – June 30, 2020, respectively.
COVID-19, coronavirus disease 2019; ALI, acute limb ischemia; CLTI, chronic limb-threatening ischemia.
Discussion
In the largest vascular registry experience published to date, including nearly 60,000 patients, we describe trends in demographic and clinical characteristics, procedural volumes, and associated clinical outcomes among patients undergoing endovascular PVI, supra-inguinal bypass, infra-inguinal bypass, and amputation before, during, and immediately following the first North American COVID-19 pandemic surge. The major observations include: (1) whereas procedure volumes for ALI remained stable over time, those for claudication and CLTI fell sharply during surge months, but rebounded quickly thereafter; (2) patients undergoing procedures during the surge were sicker, in that they had a higher prevalence of diabetes mellitus, heart failure, dialysis requirement, CLTI, and ALI than before or after the surge; and (3) despite the higher acuity of patients undergoing procedures during the pandemic surge, the unadjusted risk of most in-hospital outcomes, stratified by presenting symptom, remained stable during those months. Nevertheless, adjusted in-hospital mortality was numerically higher for those presenting with claudication and ALI during the surge and significantly higher during postsurge months.
Small, single-center studies reflecting changes in vascular practice during the COVID-19 pandemic have been published from Europe,16–18 South America, 19 and Asia, 20 as well as from some US states. 21 Most have described sharp reductions in volume, including for urgent and emergent procedures, and variable reductions in the number of procedures performed for CLTI and ALI during and following the pandemic surge; some have observed a greater degree of tissue loss at the time of presentation, and increases in the number or extent of amputations performed.16,20 Survey data from US vascular specialists have reported similar declines in procedural volume during and after the COVID-19 surge, later-stage presentation in those with CLTI, and performance of more extensive amputations than before the surge.22,23 However, no large-scale vascular registry data have been published to date. The present registry study is several orders of magnitude greater in size than that of other published reports, provides a more representative picture of the collective North American vascular experience, and affords a degree of precision not possible in these smaller studies. We observed a reduction by nearly one-half in supra-inguinal bypass, approximately one-third in PVI and infra-inguinal bypass, and nearly one-fifth for amputation procedures during as compared with before the surge. Furthermore, procedures performed for claudication fell by ~70% and those for CLTI by ~15% during versus before the surge.
The significant reduction in elective procedures performed during and following the pandemic surge was expected given governmental mandates and professional society recommendations that called for cessation or postponement of such procedures.24–26 Similarly, that the remaining patients undergoing urgent or emergent vascular procedures were sicker was not surprising. However, the significant reduction in procedures performed for CLTI and the observation that a greater proportion of these patients who underwent amputation had uncontrolled infection at the time of their procedure is of great concern and may suggest that patients delayed seeking care and/or were faced with other barriers that restricted access to the health care system. Similar phenomena have been observed with respect to other acute cardiovascular conditions such as ST-elevation acute MI and acute ischemic stroke.5,27–29 Although it is reassuring that in-hospital outcomes following procedures for CLTI during and after the surge were no different than those observed before the surge, the included registry modules do not capture outcomes in patients with vascular disease who have not undergone procedures. It remains possible that patients with CLTI who did not undergo revascularization or amputation fared worse. The new VQI Vascular Medicine Consult Registry may offer insight into this patient cohort during future pandemic surges. 30 Additionally, that the adjusted mortality rate was higher for patients presenting with claudication and ALI during postsurge than presurge months is of concern; whether this observation reflects our inability to completely adjust for patient comorbidities or is indicative of systematic differences in care, will require further study. Finally, while the numbers of patients undergoing procedures for ALI during and following the surge remained unchanged, thrombotic complications occur more often in patients with COVID-19,31,32 and it remains possible that there were additional patients with ALI in the community who, left untreated, suffered worse outcomes, including death, as a consequence.33,34
Our study provides a window into the magnitude of COVID-19-related disruption to the US and Canadian vascular care infrastructure. Although it is concerning that procedural volumes fell sharply and quickly for these high-risk patients, that they rebounded immediately following the surge demonstrates that centers can recover swiftly and offers hope that they may do so again during future pandemic surges. Nevertheless, considerable public health efforts that encourage patients with CLTI to seek care immediately will be needed moving forward; the nearly 700 patients with CLTI (likely an underestimate of the national incidence) who failed to present for life and limb-saving revascularization procedures during April and May 2020 may have lost the opportunity to do so. Likewise, initiatives that maintain access for urgent and emergent evaluation, whether through telemedicine, urgent care, emergency department or inpatient consultative services, will be critical if we are to avoid compromising vascular care in the future.
Study limitations
First, our analysis of outcomes was limited to in-hospital events, and as such may not completely characterize the impact of the pandemic on lower-extremity revascularization and amputation procedures; future studies will need to evaluate the longer-term outcome in this patient cohort. Second, our data may not reflect the experience at lower volume sites as these were excluded for statistical purposes. Third, although data on US and Canadian geographic region were available, we were unable to adjust for differences in the timing or magnitude of local community COVID-19 spread as investigators are blinded to precise hospital locations. Nevertheless, data on procedure volume by month suggest that most declines occurred during the ‘surge’ months as defined herein. Fourth, centers may create registry records at any time following the procedure; delays in doing so might disproportionately impact procedure volume estimates during the latter part of our study period as there would have been less opportunity to enter procedural data. When we restricted our analysis to records created within 60 days of the procedure, the estimated volume reduction during the surge was unchanged. However, we observed an even greater rebound in volume following the surge than observed in the overall unrestricted cohort. Although it is possible that delays in reporting occurred due to reallocation of resources and personnel during the pandemic, this would not be expected to influence our results as data are obtained through chart abstraction and do not rely on recall. Fifth, procedures performed at centers that voluntarily participate in the VQI may not represent those performed at centers who participate in other smaller registries or in no registry at all.35,36
Finally, this analysis was confined to the initial COVID-19 pandemic surge; future VQI analyses should focus on treatment and outcome differences during subsequent pandemic waves and following pandemic resolution. Additionally, while more granular data on hospital location would have allowed for a more in-depth analysis of regional differences in treatment and outcomes, these are not provided to VQI investigators in order to protect patient identity; such analyses would be useful if feasible using other datasets.
Conclusions
The North American COVID-19 pandemic surge was associated with a rapid and significant decline in endovascular and open surgical lower-extremity arterial revascularization and amputation procedures. Although patients who underwent these procedures during surge months had more advanced disease, in-hospital outcomes remained largely unchanged. Efforts to better understand the increased mortality observed during postsurge months in patients with claudication and ALI are needed. Likewise, interventions to improve access for patients with CLTI should be undertaken during future COVID-19 pandemic surges.
Supplemental Material
sj-pdf-1-vmj-10.1177_1358863X211021918 – Supplemental material for North American lower-extremity revascularization and amputation during COVID-19: Observations from the Vascular Quality Initiative
Supplemental material, sj-pdf-1-vmj-10.1177_1358863X211021918 for North American lower-extremity revascularization and amputation during COVID-19: Observations from the Vascular Quality Initiative by Jun-Yang Lou, Kevin F Kennedy, Matthew T Menard, J Dawn Abbott, Eric A Secemsky, Philip P Goodney, Marwan Saad, Peter A Soukas, Omar N Hyder and Herbert D Aronow in Vascular Medicine
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
Kevin Kennedy was a paid statistical consultant. The other authors received no financial support for the research, authorship, and/or publication of this article.
Supplementary material
The supplementary material is available online with the article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
