STRENGTHS AND LIMITATIONS OF THIS STUDY

A multicentre study with a sufficient sample size is used to increase representativeness.

The study achieved a high response rate, which helps reduce response bias.

This study only looked at symptoms of venous thromboembolism, resulting in an underestimation of the incidence.

One limitation of this study is that patients other than the emergency department were excluded from participation.

Background

Venous thromboembolism (VTE) constitutes pulmonary embolism (PE) and deep vein thrombosis resulting from blood clot in the vein which later disrupts blood flow and leads to sudden death.1 An estimated 10 million VTE episodes are diagnosed annually worldwide. Over half of these episodes are associated with hospital inpatient stays.2 A multinational, observational, cross-sectional study conducted in sub-Saharan African countries including Madagascar, Nigeria, Cameroon, the Democratic Republic of Congo and Namibia, to assess the prevalence of VTE risk in hospitalised patients reported that 62.3% of medical patients were at high risk of VTE.3

VTE has become a major public issue associated with significant mortality and morbidity.4 In the USA, approximately 60 000 to 100 000 people die from VTE each year.5 In Africa, 60% mortality rate is estimated from PE indicating that PE is a serious cause of death.6 Recent data show that the mortality rate from PE in the emergency department (ED) is estimated to exceed 20%.7 Acute complications can occur rapidly and unpredictably, making them difficult to diagnose. This is especially true for PE, where up to 50% of diagnosed cases of VTE are asymptomatic PE, leading to sudden death.8 Complications of acute VTE include post-thrombotic syndrome which can result in chronic limb heaviness, swelling, pain, paraesthesia and, in severe cases, venous leg ulcers, which occur in 4%–6% of survivors of proximal lower extremities.9 Chronic thromboembolic pulmonary hypertension is the other VTE complication that results from a PE which requires lifelong anticoagulation and is much more expensive to treat than other PE.

The financial burden of the health system is also significant.10 Once a patient develops VTE after hospital admission, medical service fees are the largest contributors to costs and the length of hospital stay for the need for VTE treatment is a factor related to increase hospitalisation costs.11

The annual incidence of VTE event is estimated to cost US healthcare more than US$7 billion each year.9 The healthcare costs of treating a patient with a VTE incident include the costs of treating the acute event as well as the costs of complications of VTE; costs associated with complications of acute VTE contribute to the economic burden of VTE.9

Although the incidence of VTE among medically ill patients varies substantially, the event rate has been reported to be 2%–10%.12–15 Most occurrences of VTE are associated with emergency care-seeking conditions such as sepsis (25%), single organ failure (63.6%) and multiple organ failure (49%), and accidental trauma (27%).16 Notably, the aspects of clot formation described by Virchow’s triad include blood stasis and vascular endothelial injury, and the onset of blood hypercoagulability can rapidly occur after acute illness.17 In this regard, VTE events, especially PE, can occur as soon as 24–72 hours after hospital admission.18 The results of studies reporting that prolonged stays in the emergency room commonly exceed 24 hours indicate that the ED should also be a place where at-risk populations should be assessed on admission and need appropriate intervention.

Risk stratification of VTE risk during hospital admission may assist clinicians and clinical pharmacists in determining the best and appropriate intervention before the occurrence of VTE.19 The results of a clinical trial revealed that stratifying VTE risk through the adoption of the Padua Prediction Score is associated with a 50% reduction in the incidence of VTE in medically ill patients compared with clinical judgement being employed by health professionals.20 However, in developing countries, including Ethiopia, priority is being given to acute disorders of current emergency admission cases with little emphasis on VTE risk assessment, which may subsequently result in worse outcomes and a significant mortality rate. However, the status of VTE in ED has not yet been assessed well in the study area. Hence, we aimed to assess VTE risk, incidence and associated factors among patients attending emergency wards of tertiary care hospitals in Addis Ababa city, Ethiopia.

MethodsStudy settings

The study was conducted on patients admitted to the adult emergency wards of three public hospitals in Addis Ababa city, Ethiopia: Tikur Anbesa Specialized Hospital (TASH), Addis Ababa Burn Emergency and Trauma Hospital (AaBET), and St. Paulos Hospital Millennium Medical College (SPHMMC). Out of the 14 public hospitals in Addis Ababa city, the three mentioned above are higher tertiary care and teaching hospitals in the country. They serve as referral centres for the communities of the capital city and all regions throughout the country, providing emergency care and inpatient services.

Patient and public involvement

Patients and the public were not involved in the design, conduct, reporting or dissemination plans of this research.

Study design and study period

A multicentre hospital-based prospective follow-up study design was employed at the ED of the three selected hospitals from June 1 to September 15, 2023.

Inclusion and exclusion criteria

All patients with complete documentation who were admitted to the adult emergency wards of the selected hospitals during the study period were included. Patients admitted with established VTE and already receiving treatment were excluded from the study.

Sample size and sampling method

The sample size for this study was determined using a single population formula. Since the incidence of VTE events in an emergency setting has not been established, assumptions of a 50% proportion, 95% CI, 5% precision level and a 10% non-response rate were considered. The calculated sample size was 422. A proportional allocation formula was then used to determine the sample proportion from each hospital. It was assumed that an average of 2080 patients visit EDs every month, with 800, 760 and 522 patients from TASH, AaBET and SPHMMC databases, respectively. This resulted in 162 patients from TASH, 154 patients from AaBET and 106 patients from SPHMMC, totaling 422 sample patients. The consecutive sampling method was employed for data collection.

Data collection, management, quality assurance and analysis

The data collection tool was designed to assess sociodemographic data such as age and sex. The Padua Risk Assessment Model was used to determine the risk of thromboembolism for each patient. A risk value of 1–4 was assigned for each criterion based on its contribution to the development of thromboembolism. The total risk score was calculated by adding points given for each Padua Risk Assessment parameter. A total score of 4 or more indicated a high risk for thromboembolism, while a total score of less than 4 was considered low risk.

The reliability of the tool was tested using Cronbach’s alpha, which was found to be 0.814. The Pearson correlation matrix also showed p<0.05, and the variance extracted for each item in the confirmatory factor analysis (CFA) was found to be more than 0.5, indicating the presence of convergent and discriminant validity.

VTE events were identified and recorded by emergency physicians on the medical cards of patients. Data collection was carried out by three nurses and the principal investigator after 1 day of training pretest was conducted at the ED of Yekatit 12 Hospital Medical College on 5% of the study population to ensure the clarity, simplicity and effectiveness of the data collection instrument. Necessary modifications were made before the actual data collection began.

Statistical analysis

Statistical Package for the Social Sciences (SPSS) Version 25 was used for analysing the data. Descriptive statistics were used to analyse relevant sociodemographic and clinical characteristics, and Cox regression analysis was performed to assess the associations between VTE occurrence and independent variables. The adjusted HR was used to measure the strength of associations. The p<0.05 was considered to indicate statistically significant. Kaplan-Meier curve was used to show survival outcomes. Significant association was indicated by the p value<0.05.

Operational definitionPadua Risk Assessment Model

This is the widely used tool to stratify medical patients at a different level of VTE risks based on the risk factors that exist in hospitalised patients.

Pharmacological prophylaxis

The use of anti-coagulants to prevent the development of thrombosis in those patients considered at risk for developing thrombosis.

Physically active

According to WHO definition, physically active includes walking, cycling, wheeling, doing sports or any form of non-motorised activity.

Venous thromboembolism incidence

The occurrence of VTE at any time after patient’s emergency admission where the diagnosis is recorded by physicians or proven from the result of the Doppler ultrasound attached to patients’ medical record.

ResultsSociodemographic characteristics

Out of a total of 422 participants recruited for the study, 218 patients (51.7%) were male. The mean (±SD) age of the patients was 49.36 (±17.12) years with a range of 19–97 years. More than half (52.1%) of the emergency attendants in the three study settings were not residents of Addis Ababa city. About 214 patients (50.7%) had community-based health insurance for medical expenses. Nearly one-third of the patients (130, 30.8%) were unable to perform physical activity due to chronic medical problems and ageing (table 1).

Table 1

Sociodemographic characteristics of the study participants at EDs of AaBET, TASH and SPHMMC from June to September 2023 (n=422)

Approximately two-thirds (76.3%) of the patients had comorbid illnesses. The most frequently reported comorbidities were hypertension (49.5%), cancer (26.3%), heart disease (acute coronary syndrome and myocardial infarction) (24.4%) and diabetes mellitus (20.3%). The most frequent admission diagnosis was trauma (23.7%), followed by infectious disease (19.2%), sepsis and septic shock (14.7%), stroke (9.7%) and acute kidney injury (5.9%). The admitted patients stayed in the emergency ward for a minimum of 1 day and a maximum of 32 days, with an average of 8 ± 6.06 days. During the study period, 47.4% of patients admitted to the ED were discharged improved, 34.6% were transferred to the internal medicine ward, 7.8% were transferred to the ICU, 3.1% were linked to the surgical ward, 4% were linked to the oncology ward and 3.1% of patients died (table 2).

Table 2

Baseline clinical characteristics of the study participants at EDs of AaBET, TASH and SPHMMC from June to September 2023 (n=422)

Levels of venous thromboembolism risk and identified Padua Prediction Score among study participants

The minimum and maximum Padua scores were 1 and 9, respectively, with the median score of 4 points. According to the Padua Risk Prediction Score, nearly two-thirds (70.62%) of patients scored ≥4 points and were at high risk of developing VTE. Immobility problem was the most common frequent risk factor (63.5%), followed by acute infection (47.4%), trauma (29.1%) and active cancer (20.6%) (table 3).

Table 3

VTE risk factors for patients admitted to ED based on the Padua Risk Prediction Score of study participants at EDs of AaBET, TASH and SPHMMC from June to September 2023 (n=422)

Venous thromboembolism event and associated factors among study participants during the follow-up period

During their emergency stay, a total of 18 patients (4.3%) were diagnosed with VTE. Of these, 16 (3.8%) had DVT and 2 (0.5%) had PE. The median time from emergency admission to VTE diagnosis was 7 days. Multivariate Cox regression analysis showed that physical activity before admission, pharmacological prophylaxis, active cancer and infection were significantly linked to VTE events during an emergency stay.

Patients admitted with infections like sepsis were about eight times more likely to experience VTE (AHR=8.169, 95% CI 1.045 to 63.854, p=0.045) compared to other groups. This indicating a significantly higher risk. Cancer patients undergoing chemotherapy were approximately five times more likely to develop VTE compared with those without active cancer (AHR=5.133, 95% CI 1.241 to 21.093, p=0.023). However, receiving pharmacological prophylaxis was associated with an 83% decrease in VTE incidence (AHR=0.167, 95% CI 0.037 to 0.768, p=0.021), and being physically active reduced VTE events by around 33% (AHR=0.67, 95% CI 0.082 to 1.579, p=0.014 (see in table 4).

Table 4

Factors associated with VTE occurrence among study participants at EDs of AaBET, TASH and SPHMMC from June to September 2023 (n=422)

According to the Kaplan-Meier survival outcome analysis, patients who were physically active (log-rank p<0.001) and who received pharmacological prophylaxis on emergency admission (log-rank p=0.001) had a longer cumulative survival time to develop VTE. However, there was also a statistically significant difference between patients with active cancer and those without active cancer (log-rank p<0.001) in terms of survival time to develop VTE and acute infection (log-rank p<0.001). Both patients with active cancer and infection had poor survival outcomes and lower cumulative survival times to VTE events than those without cancer and infection (figure 1).

Figure 1

Survival function estimate of VTE events for receiving pharmacological prophylaxis (A), physical activity (B), acute infection (C) and active cancer (D) among patients at EDs of AaBET, TASH and SPHMMC from June to September 2023 (n=422). AaBET, Addis Ababa Burn Emergency and Trauma Hospital; EDs, emergency departments; SPHMMC, St. Paulos Hospital Millennium Medical College; TASH, Tikur Anbesa Specialized Hospital; VTE, venous thromboembolism.

Discussion

A prolonged duration of stay in the ED is a global problem. Millions of individuals access healthcare through EDs but are much less concerned about the risk of VTE. This subsequently puts patients at 25% risk for death from VTE unless appropriate action is taken for prevention.2 In lower-income countries, including the study areas, the prolonged duration of patient stays in the ED remains a challenge. This is due to the increasing number of patients in need of acute care in the ED, leading to a mismatch between service demand and patient flow towards specialised units. This hinders timely access to care and the stabilisation of patients in their respective wards, particularly due to a lack of available inpatient beds, delays in investigations and waiting for senior consultations before transferring patients. As a result, overcrowding occurs in the ED, causing patients to experience prolonged stays.21

According to previous studies, the prolonged duration of patient stay at ED was 4% in England, 72.5% in Botswana, 80% in South Africa and 91.5% in Ethiopia.22–24 This indicates the highest prevalence in lower-income countries. Our study also revealed that patients admitted to ED are staying for prolonged duration at the ED much more than expected as per recommendation by Ethiopian Hospital Services Transformation Guidelines25 with an average of 8 ±6.06 days. Many of these patients admitted to the ED for medical reasons are under a direct risk for thromboembolic events unless attention is given for appropriate interventions due to acute illness severity, older age, comorbidity, immobility and other risk factors. We examined the risk of developing VTE in emergency-admitted patients and found that nearly two-thirds (70.62%) of patients at the ED of tertiary hospitals in Addis Ababa city, Ethiopia, had a high risk of developing VTE which is greater than the results reported in previous studies conducted using the same Padua VTE risk assessment tool, which showed 51% and 46.7%.26 27 This difference might be due to variations of population character as previous finding was conducted in medical ward where risk factors for hypercoagulability could be decreased post emergency after patents haemodynamically stable.

Our study revealed that the overall incidence of VTE in emergency patients admitted was 4.3%. This high occurrence could be due to a lack of optimal VTE risk stratification for emergency-admitting patients by physicians and underuse of prophylaxis for high-risk patients, which may result in patients being inappropriately exposed to VTE events. However, the finding was relatively higher compared with the results reported from Nigeria (3.1%), India (3.2%), Italy (2.4%) and the Netherlands (3.6%)28–31 and lower than the results reported from studies conducted in medical wards at Tibebe Ghion Hospital of northern Ethiopia and Hawasa Referral Hospital (6.84%–10.6%), respectively.32 33 This slight difference might be a result of poor VTE risk assessment practices and underutilisation of early intervention through VTE prophylaxis at an emergency before transferring high-risk patients to other wards leading to increase of the rate of post-emergency VTE event later in the medical ward.34

Multivariate analysis revealed that engaging in physical activity reduces the risk of VTE by 33% compared with being inactive. Consistent with this finding, reports from Norway and Sweden showed that participating in physical activity at least once a week was associated with a 28% and 17% lower risk of developing VTE, respectively.35 36 The reason for this may be blood stasis, as physical inactivity can lead to a significant localised or generalised inflammatory response. This response results in the release of inflammatory mediators from various tissues and circulatory cells, causing extensive damage to the vascular endothelium. This damage plays a crucial role in the pathophysiology of intravascular coagulation.37

Furthermore, patients who presented at the ED with acute infection were approximately eight times more likely to develop VTE than those without infection. This finding was also consistent with a retrospective study conducted in Spain among acutely ill patients which reported that having an acute infectious disease increased the likelihood of VTE events.38 This could be explained by the fact that infection may contribute to the pathogenesis of VTE by accelerating the effects of immobility and the pathogenesis may also be related to infection-induced systemic inflammation and endothelial disruption which results in a hypercoagulable state.39 Moreover, an active inflammatory response and immobilisation lead to venous stasis and effects on lung or cardiac function and thereby blood flow and homeostasis disruption.40

There is evidence that the use of anticoagulant prophylaxis through the full implementation of the American College of Chest Physicians (ACCP) guidelines in patients at VTE risk among acutely ill patients would reduce the number of death from VTE by 25%.41 The results of multivariate analysis in our study also revealed that the use of pharmacological prophylaxis for high VTE risk patients was a significant predictor of 83% hazard risk reduction. This finding was consistent with results reported in Italy and Rwanda which showed that adequate pharmacological prophylaxis reduced the risk of VTE event at 90%.42 43 Moreover, this finding was also supported by study conducted by Peter et al which reported that omission of VTE prophylaxis up on patient’s hospital admission without obvious contraindications increased odds of VTE by 85%.44

Limitation of the study

Our study has several limitations. First, since not all patients were screened for asymptomatic VTE events and the outcome was dependent only on symptomatic events confirmed by physicians, the number of episodes could be underestimated in this study. Second, long-term follow-up in terms of the occurrence of VTE after discharge and at transition wards was not assessed. Third, the small number of VTE events (18 patients) compared with the large number of non-VTE events may affect the fitness of the regression model. However, we conducted a multicentre prospective study in the largest tertiary care hospitals in the country to provide more generalisable findings, as a representative sample of patients is available from different regions in the country.

Conclusion

The proportions of patients at VTE risk and the occurrence were found to be high among emergency-admitted patients. The absence of pharmacological prophylaxis, physical inactivity prior to emergency admission, active cancer and acute infection were found to be independent predictors for the occurrence of VTE during an ED stay.

Recommendations

Hospitals should design quality improvement strategies to improve VTE risk assessment and early intervention for the prevention of occurrence at the ED by implementing relevant risk assessment tools and VTE prevention guidelines. Further prospective studies should be conducted to assess the incidence of VTE after discharge from EDs.

Data availability statement

Data are available upon reasonable request. Not Applicable.

Ethics statementsPatient consent for publication

Consent obtained directly from patient(s).

Ethics approval

This study involves human participants. Ethical approval was obtained from the Ethical Review Board (ERC) of the School of Pharmacy, Addis Ababa University (ERB/SOP/524/15/2023), for the conduct of this study at Tikur Anbesa specialised hospital, one of their teaching hospitals. The institutional review board of St. Paul’s Hospital Millennium Medical College (Pm 23/60) approved the study for its affiliated hospitals St. Paul’s Hospital and AaBET hospital. Verbal informed consent was obtained from the study participants and/or caregivers prior to their participation in the study. For the sake of anonymity, the participant’s name and medical record number were not used during data collection, and all other personnel information was kept entirely secret throughout the study period. Patients were told the reasons of being selected to be included in the study and assured that participation would not have any influence on the right to get treatment. Patients were also told about their right to withdraw from the study at any time. Participants gave informed consent to participate in the study before taking part.

Acknowledgments

We would like to express our appreciation to the Department of Pharmacology and Clinical Pharmacy, School of Pharmacy, College of Health Science, Addis Ababa University, for giving us the chance to conduct this research. The authors also thank St. Paul Millennium Medical College, Addis Ababa Burn Emergency and Trauma Hospital, and Tikur Anbesa Specialized Hospital for cooperation and permission during patient enrolment and data collection. Our acknowledgment extends to all data collectors and healthcare professionals at the three tertiary hospitals for their friendly work.