High incidence of undiagnosed deep vein thrombosis among hospitalized people living with HIV: a cross-sectional study

Abstract

Background

See More

Deep venous thrombosis (DVT) is part of a spectrum of venous thromboembolism, which also includes pulmonary embolism. Up to 50% of hospitalized individuals who develop pulmonary embolism have DVT. Pulmonary embolism is responsible for up to 10% of hospital deaths. However, the burden and specific risk factors for DVT among inpatients living with HIV in our setting are not well established. Thromboprophylaxis, which reduces the risk of DVT, is not routinely administered to all inpatients. Understanding the burden and risk factors for DVT in this population can guide the implementation of preventive measures, identify high-risk individuals, and inform guidelines for thromboprophylaxis.

Methods

This cross-sectional study was conducted between October 2023 and April 2024 in two National Referral hospitals among hospitalized people living with HIV. A total of 186 participants were consecutively sampled and enrolled in the study. All adult inpatients living with HIV who fulfilled the eligibility criteria had a bedside Doppler ultrasound scan of the lower limbs for DVT. Relevant demographic, clinical, laboratory, and HIV-related data were obtained. Both bivariable and multivariable analyses were performed via R software.

Results

Up to 186 participants were enrolled, with a median age of 40 years (interquartile range (IQR): 34–52). The prevalence of lower limb DVT was 18.3% (34/186). The participants with DVT had a median age of 49 years (IQR: 39–56, p-value 0.045). Up to 53% (99/186) of the participants were male. The median duration of hospitalization was 5.0 days (IQR: 4.0–7.8) among those with DVT. A positive history of cigarette smoking was identified in 6.5% (12/34) of the participants and was significantly associated with the development of DVT (P-value = 0.004). TB coinfection was significantly associated with DVT, with a prevalence ratio (PR) of 2.22 (P-value = 0.007). A low CD4 (< 200) was also significantly associated with lower limb DVT (PR = 2.70, P-value = 0.003).

Conclusion

The prevalence of lower limb DVT among hospitalized people living with HIV is high (18.3%) in our setting. Older age, a positive history of smoking, a low CD4 count (< 200), and TB coinfection are significant risk factors for DVT in HIV-positive inpatients.

Background

Venous thromboembolism (VTE) is a spectrum of vascular diseases that includes DVT and pulmonary embolism. Up to 50% of hospitalized individuals who develop pulmonary embolism have DVT. Pulmonary embolism is said to be responsible for up to 10% of hospital deaths [1]. HIV is a recognized risk for VTE. The incidence of DVT in hospitalized people living with HIV is not well established in sub-Saharan Africa (SSA), especially in Uganda [2]. Most of the available data are from southern Africa, where the reported prevalence of DVT among hospitalized people living with HIV is between 11% and 13% [3,4,5]. In Uganda, there is a reported incidence of DVT among out-patients living with HIV of 2.6% and 9.1% [2, 6]. However, there is paucity of in-patient data for VTE among people living with HIV.

The incidence of VTE has been reported to be approximately 1–2% higher in people living with HIV than in those without HIV [7].

The occurrence of DVT in people living with HIV has been linked to factors related to chronic immune activation and inflammation associated with HIV [8]. Acquired deficiencies of proteins S and C in people living with HIV as well as some antiretroviral therapy (ART), particularly protease inhibitors, have also been associated with the occurrence of DVT [9]. Protein S and C (glycoproteins that support the body’s natural anticoagulation mechanism) are lowered in people living with HIV and this has been attributed to decreased synthesis, the presence of antibodies to these proteins, and elevated levels of C4BP (C4b-binding protein) [10].

Thromboprophylaxis is effective at reducing the risk of thrombosis; however, its implementation in hospitalized medical patients remains low (approximately 40%) despite international guidelines for these patients [11]. The lack of consensus regarding individual patients’ risk assessment, the relative weight of the various risk factors, the application of contraindications, and the risk-benefit profile of different prophylactic measures, partly contribute to the low implementation rates among hospitalized medical patients [11]. There are no thromboprophylaxis guidelines specifically for inpatients living with HIV. Knowledge of specific risk factors for VTE among hospitalized people living with HIV could inform local guidelines for strategic and cost-effective initiation of thromboprophylaxis.

Therefore, this study aimed to determine the incidence, sonographic findings, and factors associated with undiagnosed lower limb DVT among hospitalized people living with HIV.

Methods

Study design and study setting

This was a cross-sectional study conducted between October 2023 and April 2024 in Mulago and Kiruddu National Referral Hospitals, which are both public referral hospitals found in Uganda’s capital, Kampala. The study involved hospitalized people living with HIV who were 18 years and above, and admitted to the medical ward for 72 h or more.

The study focused on patients with undiagnosed lower limb DVT that would have been missed by the attending clinicians at admission or during the course of admission. Patients who already had a confirmed radiological diagnosis of DVT and those who were receiving anticoagulation therapy were excluded.

A sample size of 186 was obtained using the Kish Leslie (1965) formula for cross-sectional studies based on a study in South Africa by Olubanwo et al., which reported a prevalence of lower limb DVT of 12.5% among hospitalized people living with HIV. Consecutive sampling of in-patients on the ward during the enrolment period was used to recruit all eligible participants.

Both sociodemographic and clinical data, such as the duration of hospitalization, associated comorbidities, CD4 cell count, antiretroviral regimen, and complete blood count parameters, were obtained for all eligible participants.

A bedside venous Doppler ultrasound scan of both lower limbs was performed using an EDAN U60 diagnostic ultrasound scan machine. The proximal and distal deep venous systems of the lower limbs were assessed for DVT, starting with the common femoral veins distally to the calf veins. DVT was defined as sonographically detected features such as the presence of a thrombus, non-venous compressibility, and the absence of venous flow. A thrombus was categorized as either acute or chronic depending on its echogenicity, deformability/flexibility, and vein distensibility characteristics. The thrombus was classified as either proximal or distal depending on the location of the affected vein relative to the knee.

Data management and analysis

Statistical data analysis, including bivariable, and multivariable analyses, was performed via R software version 4.3.3. The incidence of DVT was determined by the proportion of those with lower limb DVT to the total number of the study participants. Logistic regression was conducted to assess the factors associated with DVT. Variables that were considered for multivariable analysis included those with a P-value less than or equal to 0.2 in the bivariable analysis. Statistical significance at the multivariable level was determined at a level of significance of less than 0.05. Interactions and confounding factors were assessed and adjusted for in the multivariable analysis before the presentation of the final model. For categorical variables, either Pearson’s chi-square test or Fisher’s exact test was used depending on the number of cell frequencies, and two-sample t-tests or Mann‒Whitney U tests were used for continuous variables depending on whether the continuous variable was normally distributed.

Ethical considerations

The study was conducted after ethical approval from the Makerere University School of Medicine Research and Ethics Committee (Mak-SOMREC). Administrative clearance was also sought from Mulago and Kiruddu National Referral Hospitals. Written informed consent was obtained from all the eligible participants before enrollment in the study.

Results

A total of 186 participants were recruited, with 99 (53.2%) male participants. The median age for all the participants was 40 (interquartile range (IQR): 34.0–52.0) years. The number of participants with lower limb DVT was 34 (18.28%); 16 (47%) bilateral DVT, 18 (52.9%) proximal and 25 (73.5%) classified as acute. The median age among those with DVT and those without DVT was 49 (IQR: 39.0–56.0) and 39 (IQR: 32.0–49.0) years respectively, (P-value − 0.015). Older participants (> 50 years) were more likely to have DVT (Prevalence ratio (PR) = 1.17 [Confidence interval (CI): 1.04–1.31], p-value: 0.009) compared to younger people. The median duration of hospitalization for those with lower limb DVT and those without DVT was similar; 5.0 (IQR: 4.0–7.8) days for those with DVT and 5.0 (IQR: 4.0–7.0) days for those without DVT (p-value: 0.598).

There was no difference in the proportion of participants on ART in patients with and without DVT (20/34; 58.8% in those with DVT and 89/152; 58.6% in those without) (p-value: 0.08). there was also no difference in ART regimens in those with and without DVT (19/20; 95% vs. 82/88; 93.2) and only 6.4% (7/109) of all the participants were on second-line treatment.

A large majority of those with DVT had a CD4 cell count of less than 200 cells/µ (26/34; 76.5%) compared to those without DVT (89/152; 58.6%) (PR: 2.67, CI 1.37–5.2, P-value: 0.004) (Tables 1 and 2).

Table 1 Sociodemographic predictors of DVT using bivariable analysis

Table 2 Modified Poisson regression model examining participants’ characteristics associated with the incidence of DVT

There was a larger proportion of participants with a positive history of smoking among those with DVT (6/34; 17.6%) compared to those without DVT (6/152; 3.9%) (PR: 3.75, CI: 1.80–7.81, p-value: 0.001) (Tables 1 and 2).

A large proportion of the participants with DVT (32 (94.1%)) had a preexisting comorbidity compared to those without DVT (60.7%), (p-value: 0.003). Tuberculosis (20 (58.8%)), cryptococcal meningitis (CCM) (7 (20.6%)), and hypertension (4 (11.8%)) were the most common comorbidities among those with DVT (Table 3). However, only TB coinfection was significantly associated with DVT (PR: 2.59, CI:1.40–4.79, p-value: 0.002).

Table 3 Association of comorbidities with the occurrence of DVT using bivariable analysis

The anatomical distribution of lower limb DVT is illustrated in Fig. 1.

Fig. 1

The anatomical distribution of lower limb DVT in the deep veins. SURAL - Sural vein, PERV– Peroneal vein, PTV– Posterior tibial vein, SFV– Superficial femoral vein, CFV– Common femoral vein, POPV– Popliteal vein, DFV– Deep femoral vein, ATV– Anterior tibial vein

Discussion

The incidence of lower limb DVT among hospitalized people living with HIV was higher in our study (18.3%) compared to previous studies done in our setting and in SSA, with a previously reported prevalence ranging from 9.1 to 13.3%. This may be attributed to the fact that we screened people who were not necessarily symptomatic and included people who are admitted and are likely to have more comorbidities compared to outpatients. Up to 52.9% of those with DVT had proximal deep vein involvement and are more likely to undergo embolization leading to pulmonary embolism associated with higher mortality [12].

Of those with DVT, 47% were found to have bilateral DVT, and this was quite a high percentage. This was probably because most of our patients were too sick with hypercoagulable states. Some studies have found that patients with bilateral DVT are more likely to have a malignancy (22.5% vs. 15.4%) and pulmonary embolism (33.8% vs. 20.8%) compared to those with unilateral DVT [13]. Unfortunately, in our study, very few of our patients had a known malignancy, and those known to have any form of VTE were not enrolled.

Similar to the general population, we found that increasing age increased the risk of DVT. In the era of highly active antiretroviral therapy (HAART) people living with HIV now live longer and therefore have a recognizable risk of thrombosis both due to ageing and the ongoing inflammation [14, 15].

Participants with a positive history of smoking were 3.75 times more likely to develop DVT than nonsmokers were. This has been attributed to the vascular endothelial damage and increased hypercoagulation markers such as plasma fibrinogen which are important contributors to the development of VTE [16, 17].

Patients with TB were twice (two times) more likely to develop DVT than those without TB. In our setting, TB is still the most common opportunistic infection among people living with HIV. This was similar to what has been observed in other sub-Saharan African studies which reported TB as an independent risk factor for DVT [3, 5, 6, 18]. TB is associated with hypercoagulable states secondary to increased proinflammatory markers, high levels of plasma fibrinogen, and impaired fibrinolysis due to decreased antithrombin III, protein C, and platelet aggregation [19, 20].

A large majority of hospitalized patients living with HIV were found to have low CD4 cell counts (< 200 cells/µL). Patients with low CD4 cell counts were almost 3 times more likely to develop DVT than were those with CD4 cell counts greater than 200 cells/µL. This is linked to immune dysregulation and increased inflammation and pro-coagulation markers in patients with low CD4 cell counts [5, 18]. A large proportion (21/34; 61.8%) of the participants with lower limb DVT had lived with HIV for a shorter period of time (median(IQR): 2 (1.0-11.3) years) compared to those without DVT (median(IQR): 3 (1.0–6.0) years) and therefore it is likely that they might not have achieved adequate immune reconstitution. This is likely to predispose them to opportunistic infections.

Conclusion

The incidence of lower limb DVT among inpatients living with HIV is high (18.3%) in our setting.

The identified factors associated with lower limb DVT among inpatients living with HIV from this study included older age, a positive history of cigarette smoking, having a comorbidity, especially TB, and a low CD4 cell count (< 200 cells/µL).

We recommend screening for DVT among inpatients living with HIV, particularly those with older age, low CD4 cell count (< 200 cells/µL), TB coinfection, and a positive smoking history. The findings of this study could inform the development of specific institutional guidelines for thromboprophylaxis in hospitalized people living with HIV in our setting.

Limitations

Our study had some limitations. We were only able to conduct a Doppler ultrasound on the lower limbs which may underestimate the burden of thrombosis since it may occur at other sites including the upper limbs and higher up in the common iliac vein which may be difficult to access during standard lower limb Doppler ultrasound. Additionally, the consecutive sampling technique lacked randomness in participant selection.