Introduction

During periods of social and political instability caused by natural disasters, public health crises, and wars, health care delivery is frequently disrupted, affecting the quality of care and leading to poorer clinical outcomes. For instance, during the COVID-19 pandemic numerous countries implemented social containment measures, commonly referred to as lockdowns, to mitigate the virus transmission. These measures may have unintentionally led to delays in patients seeking emergency care, resulting in a decline in cardiovascular admissions, notably for ST-segment elevation myocardial infarction (STEMI).1-4 A failure to promptly perform revascularization in patients presenting with STEMI is believed to have contributed to increased death rate from acute coronary syndrome (ACS), heart failure, and out-of-hospital cardiac arrest observed in numerous health care systems during the pandemic.5-7 In addition, COVID-19 itself could influence the pathophysiology of ACS by increasing the risk for its thromboembolic complications.8 Similarly, natural disasters, such as the Katrina hurricane in the United States, have profoundly affected local health care, impacting not only acute illnesses9 but also influencing future cardiovascular diseases,10 and management of chronic conditions.11 Similar observations have been documented following the Kobe earthquake in Japan.12

On February 24, 2022, Russia’s invasion of Ukraine triggered Europe’s largest refugee crisis since the World War II, with an estimated 8 million people displaced within Ukraine by late May, and 12.5 million Ukrainians crossing the border of Ukraine as of November 15, 2022.13 A majority of these refugees initially sought refuge in neighboring countries to the west of Ukraine, including Poland, Slovakia, Hungary, Romania, and Moldova, resulting in over 1.5 million refugees displaced into Poland.13 Modelling estimates indicate that Ukrainian individuals may then have comprised between 15% and 30% of the population in several major Polish cities near the border.14 The impact of this unprecedented refugee displacement on health care delivery in Poland, particularly in cities closest to the Ukrainian border, remains unclear.15-17

Using STEMI as a model for acute cardiovascular care delivery, our aim was to investigate how the Russian invasion of Ukraine has indirectly affected STEMI pathways, care delivery, and clinical outcomes in Poland. Furthermore, we explored whether this impact disproportionately affected the border regions closest to Ukraine, where the population displacement has been most significant.

Patients and methods

Data source and study population

We utilized data from the national registry of percutaneous coronary intervention (PCI) (ORPKI), which is maintained by the Jagiellonian University (Kraków, Poland) in collaboration with the Association of Cardiovascular Interventions of the Polish Cardiac Society.18-22 This study encompassed all adult patients (aged >18 years) who underwent PCI for STEMI in health care centers across Poland from February 25, 2017 to May 24, 2022. In the primary analysis, we excluded PCI cases that involved thrombolysis (rescue PCI), constituting less than 0.5% of the data, as our primary focus was on patients primarily treated with PCI. In the secondary analysis, such cases were included. Due to the nature of the data (registry of procedures), neither ethics committee approval nor written informed consent from patients were required.

Outcomes

The primary outcomes encompassed: 1) the number of patients presenting with STEMI treated with PCI, and how this changed over time; 2) procedural fatality rate; and 3) the occurrence of procedural complications during angiography or PCI. Procedural complications were defined as a composite of events during PCI, including fatality rate, myocardial infarction (MI), no reflow, bleeding at the puncture site, cardiac arrest, allergic reaction, coronary artery perforation during PCI, or stroke or dissection during angiography.

The secondary outcomes, assessing whether quality metrics of STEMI care were compromised, involved consideration of the following: 1) prescription of a newer antiplatelet either pre-catheter laboratory admission or during angiography / PCI (ie, prasugrel or ticagrelor during pre-catheter laboratory admission or angiography / PCI, as opposed to a standard use of clopidogrel), 2) proportion of PCIs performed via radial access, and 3) the use of intravascular imaging (intravascular ultrasound or optical coherence tomography). Patients not prescribed a newer antiplatelet were excluded from the analysis, as the focus was on comparing the odds of prasugrel or ticagrelor prescription vs clopidogrel.

Time periods and regional centers

We categorized time periods as before February 24, 2022 (pre-war period) and after February 24, 2022 (during-war period), and classified the medical centers based on their geographic distance from the Polish–Ukrainian border. The distance from the border was determined as the shortest road route from the closest Polish–Ukrainian border crossing. For the primary analysis, we defined the regions as below 100 km vs over 100 km from the border. In a sensitivity analysis, we considered regions of below and over 200 km from the border. The centers within 100 km of the border were deemed most likely to be affected by the refugee influx, but the impact could have potentially extended further, hence the motivation for the sensitivity analysis.

Statistical analysis

We summarized baseline characteristics for the whole cohort and for individual regions across the pre- and during-war periods. Continuous variables were summarized using the mean and SD and compared using the t tests or analysis of variance. Categorical variables were summarized using frequencies and were compared using the χ2 test.

Missing data in covariate information were imputed using multiple imputation, creating 20 imputed datasets.23 Within the imputation models, we included all other variables, including the outcomes. Convergence of the imputation was confirmed. All the analyses outlined below were performed in each of the imputed datasets separately, before pooling the results using the Rubin rules.23

We calculated the number of procedures per month and per region across the study period. Given that the war in Ukraine started on February 24, 2022, we calculated each monthly procedure counts from 25th day of the previous month to the 24th day of the current month. To these data, we fitted negative binomial models between February 25, 2017 and December 24, 2021, with covariates of an indicator variable for the region (<⁠100 km vs >100 km for the main analysis, and <⁠200 km vs >200 km for the sensitivity analysis), calendar time (both continuous and as a factor variable of month to capture seasonality), an indicator variable for COVID-19 (being 1 for dates from February 24, 2020 onward, and 0 otherwise), and adjustment for the number of hospitals per region. Using these models, we then predicted the expected number of procedures per month per region, from December 25, 2021 until May 24, 2022. The predicted PCI volume was then compared with the observed monthly volume.

Mixed-effect logistic regression models with random effects per hospital were used to explore the associations between the beginning of the war in Ukraine and our patient-level outcomes of interest. All models included an indicator variable for the region (<⁠100 km vs >100 km for the main analysis, and <⁠200 km vs >200 km for the sensitivity analysis), an indicator differentiating between observations prior to February 24, 2022 (pre-war period) and after February 24, 2022 (during-war period), and the interaction between these variables. Time was included in the models as the number of months from February 25, 2017 (first day of the dataset), along with an indicator variable for COVID-19 (being 1 for the dates from February 24, 2020 onward, and 0 otherwise). We also investigated time and region indicator to examine pre-war differences in temporal changes in outcomes. The models were adjusted for the variables listed in Supplementary material, Table S1, through a propensity score (propensity for region <⁠100 km vs >100 km). For each outcome, we calculated the odds ratio (OR) with 95% CI comparing the outcomes across pre- and during-war periods, by region (<⁠100 km vs >100 km), and then tested for interactions between the region and the period (see Supplementary material, Methodology for more details).

The measure of statistical significance was set to a P value below 0.05. All analyses were undertaken in R software version 4.2.0,24 along with the “tidyverse”,25 “mice,” and “lme4” packages26 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Out of a total of 90 793 STEMI procedures included in the ORPKI registry within the study period, 9 were performed in patients younger than 18 years and further 214 were duplicate cases, all of which were excluded. For the primary analysis, another 455 procedures were excluded, as the patients received thrombolysis during angiography or PCI, resulting in 90 115 procedures included in this analysis. For the secondary analysis, thrombolysis cases were included, with a sample size of 90 570 procedures over the study period. The primary analysis dataset included a total of 162 hospitals, 8 of which were located less than 100 km from the border.

Table 1 provides the baseline characteristics of the primary analysis cohort, both overall and by region-time period combinations, with Supplementary material, Table S2 summarizing the same data for the secondary analysis cohort that included thrombolysis cases. In the primary analysis cohort, the mean (SD) age was 65 (12) years, and 68% of the patients were men. Overall, 17% had diabetes, 31% were current or previous smokers, and 12% had a history of PCI. As many as 4.4% of the cohort patients presented with cardiac arrest at baseline. A majority (81%) of the whole cohort underwent angiography and PCI via the radial access.

Table 1. Baseline characteristics of the primary percutaneous coronary intervention cohort, excluding thrombolysis cases, the main analysis

Parameter

Overall (n = 90 115)

Region <⁠100 km

Region >100 km

Pre-war (n = 3854)

During war (n = 154)

P value

Pre-war (n = 82 736)

During war (n = 3371)

P value

Age, mean (SD)

65 (12)

66 (12)

67 (11)

0.53

65 (12)

66 (12)

0.02

Sex

Women

28 636 (32)

1222 (32)

52 (34)

0.61

26 287 (32)

1075 (32)

0.86

Men

61 237 (68)

2618 (68)

102 (66)

56 233 (68)

2284 (68)

Missing data

242

14

0

216

12

Diabetes

15 751 (17)

606 (16)

23 (15)

0.79

14 528 (18)

594 (18)

0.93

Previous stroke

2786 (3.1)

118 (3.1)

6 (3.9)

0.48

2577 (3.1)

85 (2.5)

0.05

Previous MI

11 278 (13)

465 (12)

20 (13)

0.73

10 360 (13)

433 (13)

0.58

Previous PCI

11 100 (12)

448 (12)

22 (14)

0.31

10 173 (12)

457 (14)

0.03

Previous CABG

1522 (1.7)

48 (1.2)

1 (0.6)

>0.99

1419 (1.7)

54 (1.6)

0.62

Previous smoker

28 023 (31)

1023 (27)

32 (21)

0.11

25 974 (31)

994 (29)

0.02

Hypertension

52 609 (58)

2613 (68)

99 (64)

0.36

47 967 (58)

1930 (57)

0.4

Kidney disease

3059 (3.4)

153 (4)

4 (2.6)

0.39

2817 (3.4)

85 (2.5)

0.01

COPD

1998 (2.2)

95 (2.5)

1 (0.6)

0.18

1834 (2.2)

68 (2)

0.44

Cardiac arrest at baseline

3816 (4.2)

181 (4.7)

8 (5.2)

0.77

3508 (4.2)

119 (3.5)

0.04

Killip class

I

62 654 (83)

2639 (87)

97 (84)

0.21

57 632 (83)

2286 (83)

0.61

II

7865 (10)

227 (7.4)

8 (7)

7328 (11)

302 (11)

III

2415 (3.2)

65 (2.1)

6 (5.2)

2265 (3.2)

79 (2.9)

IV

2726 (3.6)

116 (3.8)

4 (3.5)

2510 (3.6)

96 (3.5)

Missing data

14 455

807

39

13 001

608

ASA, pre-catheter laboratory admission

66 321 (74)

2891 (75)

112 (73)

0.52

60 766 (73)

2552 (76)

<⁠0.001

ASA, during angiography or PCI

72 368 (80)

3190 (83)

131 (85)

0.46

66 298 (80)

2749 (82)

0.04

UFH, pre-catheter laboratory admission

48 679 (54)

2590 (67)

103 (67)

0.93

44 142 (53)

1844 (55)

0.12

UFH, during angiography or PCI

79 024 (88)

3629 (94)

152 (99)

0.02

72 310 (87)

2933 (87)

0.50

LMWH, pre-catheter laboratory admission

1755 (1.9)

41 (1.1)

3 (1.9)

0.24

1517 (1.8)

194 (5.8)

<⁠0.001

LMWH, during angiography or PCI

3651 (4.1)

53 (1.4)

3 (1.9)

0.48

3391 (4.1)

204 (6.1)

<⁠0.001

GPI IIb/IIIa

24 349 (27)

888 (23)

44 (29)

0.11

22 619 (27)

798 (24)

<⁠0.001

Results of angiography

LMCA disease

6539 (7.3)

251 (6.5)

13 (8.4)

0.61

6017 (7.3)

258 (7.7)

0.7

Multivessel disease

43 394 (48)

1895 (49)

76 (49)

39 812 (48)

1611 (48)

Single-vessel disease

40 182 (45)

1708 (44)

65 (42)

36 907 (45)

1502 (45)

Bivalirudin

610 (0.7)

11 (0.3)

0

>0.99

579 (0.7)

20 (0.6)

0.47

FFR

189 (0.2)

9 (0.2)

0

>0.99

172 (0.2)

8 (0.2)

0.71

Intravascular imaging

1531 (1.7)

99 (2.6)

5 (3.2)

0.6

1302 (1.6)

125 (3.7)

<⁠0.001

Aspiration thrombectomy

9544 (11)

609 (16)

25 (16)

0.89

8586 (10)

324 (9.6)

0.15

Rotablation

103 (0.1)

2 (<⁠0.1)

0

>0.99

98 (0.1)

3 (<⁠0.1)

>0.99

Access site

Femoral – any side

15 170 (17)

943 (24)

15 (9.7)

<⁠0.001

13 863 (17)

349 (10)

<⁠0.001

Radial – any side

72 910 (81)

2813 (73)

138 (90)

67 012 (81)

2947 (87)

Radial angiofemoral PCI

1041 (1.2)

67 (1.7)

1 (0.6)

940 (1.1)

33 (1)

Femoral angioradial PCI

315 (0.3)

16 (0.4)

0

292 (0.4)

7 (0.2)

Other

679 (0.8)

15 (0.4)

0

629 (0.8)

35 (1)

Total contrast

166 (70)

193 (77)

180 (67)

0.03

165 (70)

156 (63)

<⁠0.001

Missing data

2651

110

0

2431

110

Total radiation dose, mGy

834 (731)

892 (804)

650 (620)

<⁠0.001

836 (730)

730 (645)

<⁠0.001

Missing data

2464

117

0

2235

112

Infarct-related artery

RCA

36 833 (41)

1602 (42)

62 (40)

0.75

33 850 (41)

1319 (39)

0.04

LMCA

2314 (2.6)

86 (2.2)

7 (4.5)

0.09

2129 (2.6)

92 (2.7)

0.58

First / second diagonal

4190 (4.6)

189 (4.9)

6 (3.9)

0.57

3857 (4.7)

138 (4.1)

0.12

Circumflex

12 164 (13)

526 (14)

19 (12)

0.64

11 137 (13)

482 (14)

0.16

First / second / third obtuse marginal

3775 (4.2)

147 (3.8)

7 (4.5)

0.64

3490 (4.2)

131 (3.9)

0.35

Ramus intermedius

924 (1)

36 (0.9)

4 (2.6)

0.07

851 (1)

33 (1)

0.78

Direct transport

24 093 (27)

856 (22)

63 (41)

<⁠0.001

22 209 (27)

965 (29)

0.02

Data are presented as number and percentage unless indicated otherwise.

Abbreviations: ASA, acetylsalicylic acid; CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; FFR, fractional flow reserve; GPI, glycoprotein; LAD, left anterior descending coronary artery; LMCA, left main coronary artery; LMWH, low-molecular-weight heparin; MI, myocardial infarction; PCI, percutaneous coronary intervention; RCA, right coronary artery; UFH, unfractionated heparin

In general, we found that the baseline characteristics within each region were similar across the time periods before and during the war (Table 1; Supplementary material, Tables S2–S4). However, in the centers located over 100 km from the border, the proportion of patients with previous PCI, previous smokers, and those with kidney disease were significantly different before and after February 24, 2022; such differences were not significant in the centers located less than 100 km from the border (Table 1). Interestingly, the proportion of patients who were directly transported increased substantially during the war for the centers within 100 km of the border (from 22% before February 24, 2022 to 41% after this date).

Percutaneous coronary intervention volume

Supplementary material, Figure S1 shows the number of procedures per month for the whole cohort. We observed a steady decrease in the overall number of procedures with time, which became evident rapidly post–COVID-19. Overall, there was little change upon the war starting in Ukraine, but this varied across regions. Specifically, when comparing the predicted and observed monthly volume per region (based on historic trends), the actual average number of procedures per month per region was similar to that predicted for the centers located over 100 km from the border (Figure 1). However, in the centers situated less than 100 km from the border, the average number of procedures was higher than expected during the month immediately following the beginning of the war in Ukraine (by an estimated median of 15 [interquartile range, 11–19] more procedures), as compared with predicted levels based on historic trends (Figure 1). The value returned to the expected levels after 30 days. Similar results were obtained in the sensitivity analysis covering the regions below and over 200 km from the border, and in the secondary analysis that included thrombolysis cases.

Figure 1. Observed (grey) and predicted (blue) number of procedures per month per region. The dotted line represents the beginning of the COVID-19 pandemic, and the dashed line represents the first day of the war in Ukraine.

Procedural outcomes and parameters of care

In the region closer to the border (<⁠100 km), the procedural fatality rate approximately doubled after the war began, but this increase was found insignificant after multivariable adjustment (OR, 2.24; 95% CI, 0.78–6.48; Table 2). In the region located further from the border (>100 km), the fatality rate was similar before and during the war (OR, 1.01; 95% CI, 0.66–1.55), and there was no evidence of a difference-in-difference for the odds of procedural fatality rate between the regions (P value for the interaction was 0.17; Table 2). This was also observed in the sensitivity analysis for the regions located less and more than 200 km from the border, and in the secondary analysis cohort including thrombolysis cases (Supplementary material, Tables S5–S7).

Table 2. Associations between the beginning of the Russian invasion of Ukraine and clinical outcomes in the primary percutaneous coronary intervention cohort, excluding thrombolysis cases, for the regions within and over 100 km from the Polish–Ukrainian border

Outcomes

Region <⁠100 km

Region >100 km

P value for interactionb

Pre-war (n = 3854)

During war (n = 154)

aOR (95% CI)a

Pre-war (n = 82 736)

During war (n = 3371)

aOR (95% CI)a

Procedural fatality rate

47 (1.22)

5 (3.25)

2.24 (0.78–6.48)

825 (1)

24 (0.71)

1.01 (0.66–1.55)

0.17

PCI complications

188 (4.88)

14 (9.09)

1.10 (0.59–2.05)

3458 (4.18)

109 (3.23)

0.79 (0.64–0.97)

0.31

New antiplatelet medication

1855 (57.5)

103 (74.1)

0.73 (0.47–1.13)

31 537 (46.4)

1765 (63.9)

0.62 (0.56–0.69)

0.47

Radial access

2829 (73.4)

138 (89.6)

1.15 (0.66–2.01)

67 304 (81.4)

2954 (87.6)

0.94 (0.84–1.06)

0.49

Imaging (IVUS/OCT)

99 (2.57)

5 (3.25)

0.66 (0.25–1.73)

1302 (1.57)

125 (3.71)

0.86 (0.69–1.06)

0.61

Data are presented as number and percentage.

a Odds ratios were adjusted for time since the beginning of the study, COVID-19, propensity score (see Supplementary material, Table S1), and hospital-level random effect; see Supplementary material, Methodology.

b P value for interaction tested the interaction between the pre- and during-war period indicator variable with region indicator variable, to determine whether the outcomes between the pre- and during-war periods differed by region.

Abbreviations: aOR, adjusted odds ratio; IVUS, intravascular ultrasound; OCT, optical coherence tomography; others, see Table 1

Upon examining quality of care indicators, we found no differences for the odds of PCI complications before and during the war in the regions closer to and further from the border, and also no evidence of a difference-in-difference across the regions (Table 2). Our findings were similar for the prescription of newer antiplatelet medication, radial access, and imaging outcomes.

Discussion

Our analysis revealed that following the Russian invasion of Ukraine and the subsequent displacement of 1.5 million refugees into Poland, there has been an increase in revascularization procedures for STEMI, disproportionately impacting the centers within 100 km of the border. The rise in numbers has been relatively modest, estimating additional 15 procedures per month within 100 km of the border. After multivariable adjustment, there was no evidence of a difference in clinical outcomes.

The increase in the number of PCIs has been relatively modest as well, with estimated 15 extra procedures within 100 km of the Polish–Ukrainian border. This rise may be below expectations considering the sudden displacement of over 1.5 million refugees to the regions close to the border. There are several possible explanations. First, a majority of the refugees were women with young children, as men under the age of 60 were prohibited from leaving Ukraine as part of the country’s enlistment regulations. Second, our analysis was based on the primary PCI activity using the Polish national PCI registry as a surrogate for STEMI admissions. It is possible that during this period, a greater proportion of patients with STEMI presented outside the 12-hour timeframe in which primary PCI is believed to be effective. Although possible, this is unlikely, as we did not observe a difference in either time from pain-to-first contact or time from pain-to-angiography / balloon angioplasty. Furthermore, it is reassuring to note that the increase in primary PCI number did not create a bottleneck in the system, causing delays in pain-to-balloon times. A third possibility is that the refugees who experienced STEMI may not have sought medical assistance due to unfamiliarity with the medical system in Poland or a lack of access to medical facilities.

The rise in STEMI admissions observed in the current study in the border regions (<⁠100 km) may not have resulted merely from the population enlargement due to the refugee displacement. Prior research showed that civil unrests and natural disasters are associated with higher incidence of cardiovascular events. For instance, following the Northridge earthquake in the Los Angeles area in January 1994, a postal survey of more than 100 hospitals in the region indicated that admissions for acute MI (AMI) increased from 149 in the week before to 201 in the week after the earthquake.27 Additionally, the number of sudden deaths from cardiac causes increased from an average of 4.6 per day in the week before to 24 on the day of the earthquake.28 This surge in incident MI following major events might last for extended time periods, as heart attack centers reported a 3-fold increase in the incidence of AMI 2 years after the Katrina hurricane in the United States.29

Emotional stress during natural disasters and in war zones may elevate the incidence of acute coronary syndromes through various mechanisms, including excessive sympathetic nervous system activation, glycemic control, exacerbation of coronary artery atherosclerosis, transient endothelial dysfunction or necrosis, and increase in platelet aggregability.29,30 Other prospective studies have suggested that the association between psychological distress and cardiovascular disease risk could be largely explained by behaviors such as smoking and alcohol intake.31 Alternatively, it may be related to social factors, such as missed medications, changed diet, poor living conditions, and stress due to commotion and crime following the event.32

We observed an increase in in-hospital fatality rate in the centers within 100 km of the border following the Russian invasion of Ukraine, although its significance diminished when adjusting for differences in baseline covariates. This is reassuring, since we did not identify an impact of the war on the patient care parameters, such as pain-to-balloon time or prescription of newer antiplatelet agents. However, considering the relatively large confidence intervals for the odds of mortality before and during the war (OR, 2.22; 95% CI, 0.77–6.42), we cannot entirely rule out the effect on fatality rate. Limited total number of cases in the centers within 100 km of the border could have affected the statistical power to detect differences. The challenge to the health care system may still have resulted in lowering the quality of care delivered in these regions, which was not captured by the ORPKI dataset. These unobserved factors, known to impact clinical outcomes, include admission to coronary care, prescription of statins, β-blockers, and angiotensin-converting enzyme inhibitors.33

Our study did not identify a population shift toward a worse risk factor profile following the influx of refugees, as the average age of the patients treated at the border centers did not change following the invasion. The risk factor profiles were similar before and during the invasion, with comparable proportions of Killip class III/IV presentations. It is likely that the impact on the health care was much greater than reported in our study, with challenges involving prescription of longer-term antiplatelet agents to a population of migrants who may lack a stable access to medical care. Additionally, challenges may arise in secondary prevention, including optimal blood pressure control, lipid management, cardiac rehabilitation, and altered availability of usual medication due to supply chain issues. Furthermore, as highlighted previously, the impact of natural disasters, wars, and other social upheavals may affect cardiovascular health several years post the index event.

Previous studies examining the impact of conflict events on ACS reported variable findings. One study indicated that the incidence of admissions for AMI increased during the first 5 days of air raids (incident rate ratio, 2.43; 95% CI, 1.23–4.26), and other studies noted a significant rise in the number of patients with wartime ACS in Bosnia and Herzegovina as compared with the pre-war period.34-36 In contrast, a study focusing on coronary care unit admissions at 8 centers in New York City after the September 11 terrorist attacks on the World Trade Center did not find a significant change in the number of admissions for ACS or chest pain in the week following the attacks.37 Our study contributes to this body of literature by evaluating the parameters of patient care and clinical outcomes, and assessing whether they have been impacted by population shifts, particularly in relation to the geographic proximity to the conflict zone. Despite papers highlighting correlations between humanitarian crises and cardiovascular morbidity and mortality, much of the planning of humanitarian responses has concentrated on communicable diseases. The findings of our analysis suggest that the consideration of acute cardiovascular diseases such as AMI in disaster planning and response efforts may help mitigate cardiovascular morbidity and mortality.

Limitations

Our analysis has several limitations. First, it utilizes the ORPKI national PCI registry, which only captures the STEMI admissions that proceed to primary PCI, not the total number of STEMI presentations. Second, the ORPKI dataset does not link to longer-term outcomes, and therefore cannot explain if the worse clinical outcomes following the Russian invasion of Ukraine translate into worse longer-term outcomes in the border regions. Third, while we could model the changes in the number of procedures per month, we did not have data on which PCI procedures were performed in the Ukrainian refugees specifically. Fourth, limited number of PCI cases in the regions within 100 km of the borders may affect the statistical power of the study. Finally, our results should not be interpreted as causal, although we have used propensity score analyses to adjust our estimates for the observed confounders.

Conclusions

In conclusion, our analysis of the national Polish PCI registry reveals that, following the Russian invasion of Ukraine and the displacement of 1.5 million people into Poland, there was only a modest and temporary increase in primary PCI activity, predominantly in the centers located within 100 km of the Polish–Ukrainian border. However, there was no significant impact on in-hospital fatality rate. Our findings suggest that planning for both monitoring and management of cardiac diseases should be prioritized in conflict zones and scenes of natural disasters.