Prevalence and incidence of venous thromboembolism in geriatric patients admitted to long-term care hospitals

The risk of venous thromboembolism (VTE) increases with age. However, the risk of VTE in the setting of long-term care hospitals is understudied. Our objective was to provide data on the prevalence and incidence of VTE in older adults admitted to long-term care hospitals. In this retrospective cohort study, we collected data about chronically ill and multimorbid patients aged 65 years and older from two long-term care hospitals. The primary endpoint of this study was the lifetime prevalence of VTE, and the secondary endpoint was VTE incidence during residency in long-term care hospitals. We analysed data from 1148 patients with a mean age of 84.1 ± 7.9 years, of whom 74.2% were women. The lifetime prevalence of VTE at baseline was 9.6% (95% CI 7.9–11.4). Cumulative incidence of VTE at 1, 2, and 3 years from baseline was estimated at 3.5% (95% CI 2.5–4.7), 4.2% (95% CI 3.1–5.5), and 5.4% (95% CI 4.1–7.0), respectively. Overall, the incidence rate of VTE in our study was 2.82 (95% CI 2.18–3.66) per 100 person-years. The study indicated a considerably high lifetime prevalence and incidence of VTE during residence in long-term care hospital settings, requiring further evaluation in larger prospective studies.


Patient population
We collected data from chronically ill and multimorbid patients who were admitted to two long-term care hospitals, namely Haus der Barmherzigkeit Tokiostrasse and Haus der Barmherzigkeit Seeboeckgasse in Vienna, Austria.The data collection period ranged from 1 January 2014 to 31 October 2017.For our cohort study, patients 65 years or older with varying levels of care dependency and underlying chronic medical conditions were eligible.Usually, patients admitted to a long-term care hospital are those with multimorbidity (e.g., COPD, heart failure, frailty, stroke, diabetes) requiring continuous medical care at a level that cannot be provided in a nursing home.We retrieved data from electronic medical records about comorbidities, geriatric assessment, medication, laboratory results, date of admission, discharge, and death.The Charlson Comorbidity Index 20 was calculated for each patient based on available comorbidity data.For inpatients on 1 January 2014, comorbidities and endpoints were considered as past medical history if dated on or before this date.For patients admitted between 1 January 2014 and 31 October 31 2017, diagnoses were considered as past medical history if dated prior to admission or up to five days after admission.We identified prescribed oral anticoagulants (non-vitamin K oral anticoagulants [NOAC], vitamin K antagonist [VKA]), parenteral anticoagulants (low-molecular-weightheparin [LMWH], fondaparinux), and platelet inhibitors according to the Anatomical Therapeutic Chemical classification system.Medication was considered present at baseline if it was prescribed up to five days after 1 January 2014 or after admission.The findings on the prevalence and incidence of atrial fibrillation in this cohort have been previously published 21 .

Endpoints
The primary endpoint of this study was the lifetime prevalence of VTE at baseline, and the secondary endpoint was the incidence of VTE during residence in long-term care hospitals.VTE was defined as any provoked (i.e., presence of any obvious cause or risk factor) or unprovoked (i.e., absence of any obvious cause or risk factor) pelvic or lower extremity DVT and/or PE.Using International Classification of Disease, 10 th Revision (ICD-10) codes assigned by treating physicians and/or free-text information from electronic medical records, we identified patients who experienced pelvic or lower extremity DVT (ICD-10 codes I80.1, I80.2, I80.3) and/or PE (ICD-10 code I26.-).When necessary, two clinicians involved in this study manually reviewed medical records to clarify the diagnoses and dates of outcome events.

Statistical analysis
Comprehensive descriptive statistics for baseline and endpoint variables were performed.For continuous data, we computed means and standard deviations if approximately normally distributed; otherwise, the median and the 25th and 75th percentiles were calculated.Categorical data were summarised as absolute and relative frequencies.We determined the lifetime prevalence of VTE at baseline (i.e., 1 January 2014 or date of admission for those hospitalised in the period from 2 January 2014 and 31 October 31 2017) with a 95% confidence interval (CI) by calculating the proportion of patients with history of VTE during their lifetime.
Regarding the incidence of VTE, we estimated the cumulative incidence and considered death from any cause as the competing risk event 22 .We calculated the time from baseline to occurrence of VTE (date of diagnosis or first documentation in the electronic medical record).If no VTE was observed, we calculated the time until discharge from a long-term care hospital, death, or the end of the follow-up period (31 October 2017).We determined incidence rates by dividing the number of VTE events by years at risk in person-time, considering the observation period from baseline to discharge, death, or the end of the follow-up period (October 31, 2017) as the person-time at risk.We performed stratified analyses for incidence rates according to history of VTE (lifetime history versus no history of VTE) and the presence of anticoagulation (oral anticoagulation versus parenteral anticoagulation versus no anticoagulation) at baseline.Statistical analysis for this study was conducted using SAS® software version 9.4 (SAS Institute Inc., Cary, NC, USA) and Stata Statistical Software: Release 17.0 (StataCorp LLC, College Station, TX, USA).

Patient characteristics
The mean age of the patients was 84.1 ± 7.9 years; 74.2% of the patients were female.More than half of the patients (51.0%) were 85 years or older, and more than 80% of patients had a Charlson Comorbidity Index below 4.Among the patients with a history of VTE, 32.7% (36 out of 110) received oral anticoagulants, and 15.5% (17 out of 110) received parenteral anticoagulants.Table 1 shows the baseline characteristics of patients with and without a history of at least one VTE.Supplementary Table S1 presents the patients' baseline characteristics by sex.

Prevalence of venous thromboembolism
At baseline, of 1,148 patients, 110 had a history of VTE during their lifetime: 60 had experienced DVT, 34 PE, and 16 both DVT and PE.Consequently, the baseline lifetime prevalence of VTE was determined to be 9.6% (95% CI 7.9-11.4).The incidence rate of VTE in the entire study population was 2.82 (95% CI 2.18-3.66)per 100 person-years.Among patients without a history of VTE, the incidence rate was 2.94 (95% CI 2.25-3.84)per 100 person-years.During the follow-up period, among the 110 patients with a history of VTE at baseline, 3 patients experienced recurrent VTE.
Results of a subgroup analysis stratified by sex are provided in the supplementary material (Tables S2-S4).

Discussion
In this retrospective cohort study of geriatric patients admitted to long-term care hospitals, we identified a high lifetime prevalence of VTE of approximately 10%, along with an overall incidence rate of 2.82 per 100 personyears during their stay in long-term care hospitals.Comparing our findings to previous studies conducted on nursing home residents, we observed a higher prevalence of VTE.For instance, a study from the US reported a history of VTE diagnosis on admission in 3.7% of all admissions across 181 nursing homes 23 .It should be noted that the patients included in this study were younger and had fewer comorbidities compared to our study population.In addition, comparability of our study is limited due to the methodological differences, e.g., the aforementioned study included patients from nursing homes (compared to long-term care hospitals in our study), and the outcome was based on Minimum Data Set (MDS) 2.0 assessments 23 .
The incidence of VTE observed in our study is not only significantly higher compared to the general population 1,2 , but it also exceeds population-wide estimates for individuals above the age of 80, which range around 0.8 VTE events per 100 person-years 14 .However, when considering studies performed in a nursing home setting, our findings demonstrate comparable incidence rates, indicating a substantial burden of VTE in geriatric patient populations.
While we were unable to find specific data from long-term care hospitals, there are previous studies that assessed VTE incidence using administrative data from one or more nursing homes and reported incidence rates ranging from 1.3 per 100 person-years 24 up to 3.7 per 100 person-years 23 .An observational cohort study conducted in the UK found rates between 0.71 and 2.48 per 100 person-years, depending on the inclusion of probable and possible VTE events 17 .Similarly, the authors of a cohort study of nursing home residents in Israel observed incidence rates between 1.4 and 1.6 per 100 person-years 25 .The most recently published study, including nursing home residents in the US, revealed an incidence rate of 3.65 per 100 person-years 18 .Proposed reasons for the variation in VTE incidence across these studies include differences in patient characteristics, setting, follow-up duration, and the availability of and access to proper diagnostic methods 23 .The higher incidence of VTE in our study compared to the general population and its similarity to rates reported in nursing home studies indicate a substantial burden of VTE in older adults admitted to long-term care hospitals.

Table 2.
Antithrombotic therapy in patients with newly diagnosed VTE during follow-up.Abbreviations: NOAC, non-vitamin K antagonist oral anticoagulant; LMWH, low-molecular-weight-heparin; N, number of patients; VKA, vitamin K antagonists; VTE, venous thromboembolism.a Patients who initially received LMWH or fondaparinux and subsequently switched to oral anticoagulation were considered in the category of prescribed oral anticoagulation.b One patient died from pulmonary embolism, so no long-term anticoagulation therapy was recorded.The high incidence of VTE observed in our study can be attributed to several factors.The geriatric population in long-term care hospitals is characterised by a high prevalence of chronic illnesses, including malignancies, which are known to increase the risk of VTE.This underlying comorbidity burden contributes significantly to the elevated incidence of VTE in this specific population.Moreover, the setting of long-term care hospitals differs in several important aspects from other healthcare institutions, such as nursing homes.Notably, long-term care hospitals provide continuous medical care by physicians, including regular ward rounds.This increased level of medical attention and close monitoring may enhance the detection rate of clinical events, including VTE, compared to other healthcare settings.
It is important to note that the immediate and accurate diagnosis of VTE has a significant impact on epidemiological findings.While a correct diagnosis is typically made promptly in symptomatic patients, the possibility of overlooking VTE in the absence of typical symptoms exists.Consequently, the incidence of VTE might have been underestimated in our study due to the presence of clinically silent thrombotic events.This hypothesis is supported by the findings of an asymptomatic VTE point prevalence of 18% in chronically ill patients who underwent systematic screening 26 .Moreover, an autopsy study revealed that PE was the cause of death in 8% of residents in a chronic care institution 27 , highlighting the high rate of missed diagnoses prior to death.These data emphasise the crucial role of VTE detection in the context of long-term care settings and highlight the necessity for increased attention regarding its correct diagnosis.
The use of prophylactic or therapeutic anticoagulation could have potentially reduced the risk of thrombosis in individual patients in our study.However, the overall frequency of antithrombotic therapy in patients with a history of VTE at baseline was relatively low.This could be explained by several factors, including the fact that some of the VTE events may have occurred long ago, leading to the discontinuation of therapy, or the VTE events were provoked by transient risk factors and treatment was performed for a limited duration.However, patients in our study also had other indications for anticoagulation, such as atrial fibrillation 21 .Finally, the variable follow-up duration among the patients in our study could have potentially affected the incidence rates observed.
Our observed decrease in VTE incidence rates among patients stratified by oral anticoagulant therapy, parenteral anticoagulation, and no anticoagulation should be interpreted with caution.Due to the small sample size and the presence of confounding factors, it does not provide reliable evidence to draw any conclusions regarding VTE prevention.The subgroup of patients receiving oral anticoagulation included individuals on therapeutic anticoagulation for reasons unrelated to VTE, such as atrial fibrillation, and the parenteral anticoagulation subgroup comprised both prophylactic and therapeutic uses.In addition, clinical decisions regarding anticoagulation and relevant information such as clinical presentation, patient preference, estimated life expectancy, and bleeding risk were not available from our data set.Finally, we assessed the use of anticoagulation at baseline and at the time point of a new VTE diagnosis during follow-up.We were not able to take into account changes in anticoagulation during follow-up as a time-dependent co-variate.
In contrast to our findings, other studies have reported significantly lower rates of anticoagulant usage, ranging from 3 25 to 6% 17 .However, it is important to note that these studies had different designs and did not specifically evaluate the incidence rates of VTE in relation to anticoagulant use.For example, the study by Gatt et al. was a historical cohort study conducted in a nursing home which did not specify the time point of medication assessment 25 .Apenteng et al. included more than 1000 nursing home patients in a prospective cohort study, in which approximately 6% of patients received anticoagulation at baseline, primarily for atrial fibrillation 17 .
While there are valuable data regarding thromboprophylaxis in hospitalised medical and surgical patients, the available evidence for long-term care residents and patients admitted to long-term care hospitals is insufficient.Furthermore, to our knowledge, no clinical prediction tool is available to aid the decision-making process.Therefore, the most appropriate approach at present is to engage in individualised, patient-centred discussions when considering thromboprophylaxis.These discussions should take into account the cumulative impact of risk factors for VTE while also carefully assessing the risk of bleeding 28,29 .Additionally, patients' values and preferences should be acknowledged in shared decision making.
Our study has several limitations that should be considered.First, it is susceptible to bias due to its retrospective design with data from electronic medical records.Second, the generalizability of our findings from two long-term care hospitals to other healthcare settings, such as nursing homes or acute care hospitals, is limited.The specific characteristics and patient populations in long-term care hospitals differ from those in other healthcare facilities.Third, our analysis regarding the usage of anticoagulants and occurrence of new VTE during follow-up is exploratory in nature.Insufficient information regarding dosage, indication, and duration of anticoagulation limits our ability to draw any conclusions.We were not able to evaluate the underlying reason for the decisions for or against anticoagulation, as this information was not available in the data set.Finally, we could not obtain information on whether the VTE events were provoked or unprovoked.
In conclusion, our study highlights the significant burden of VTE among older adults admitted to long-term care hospitals.To confirm our findings on the prevalence and incidence of VTE in long-term care hospitals, there is a need for large prospective studies.Future studies might advance our understanding of VTE in this vulnerable patient population, support physicians in making more informed decisions, and optimise the prevention of VTE in long-term care settings.

Figure 1 .
Figure 1.Study flow chart showing the number of patients with VTE at baseline and during follow-up.VTE, venous thromboembolism; N, number of patients.

Table 2
presents the antithrombotic regimens in patients who experienced VTE during the follow-up period.Most of the patients received LMWH, followed by NOAC.