Safety and efficacy of low-dose fondaparinux (1.5 mg) for the prevention of venous thromboembolism in acutely ill medical patients with renal impairment: the FONDAIR study



This article is corrected by:

  1. Errata: Corrigendum Volume 14, Issue 1, 225, Article first published online: 13 December 2015

Walter Ageno, Department of Clinical and Experimental Medicine, Research Center on Thromboembolic Diseases and Antithrombotic Therapies, University of Insubria, Varese, Italy.
Tel.: +39 332278594; fax: +39 332393640.


Summary.  Background:  Renal impairment is common, affecting around 40% of acutely ill medical patients, and is associated with an increased risk of both venous thromboembolism (VTE) and bleeding. The clinical benefit of effective thromboprophylactic strategies may be outweighed in these patients by an excessive rate of hemorrhage.

Objective:  To assess the safety and efficacy of lower prophylactic doses of fondaparinux in acutely ill medical patients with renal impairment.

Patients/Methods:  We carried out a multicenter, investigator-initiated, prospective cohort study. Patients at risk of VTE with a creatinine clearance between 20 and 50 mL min−1 were treated with fondaparinux 1.5 mg qd for a minimum of 6 to a maximum of 15 days. The primary outcome was the incidence of major bleeding; secondary outcomes were clinically relevant non-major bleeding (CRNMB) and symptomatic VTE.

Results:  We enrolled 206 patients with a mean age of 82 years, mean creatinine clearance of 33 mL min−1, and a mean Charlson co-morbidity index of 8.2. One patient had major bleeding (0.49%, 95% confidence interval [CI] 0.03–3.10), eight had CRNMB (3.88%, 95% CI 1.81–7.78) and three developed symptomatic VTE (1.46%, 0.38–4.55). Twenty-three patients (11.17%, 7.36–16.48) died. No independent predictors of bleeding were found at univariate analysis.

Conclusions:  The addition of moderate to severe renal impairment to patients with traditional risk factors for VTE identified a population of very elderly acutely ill medical patients potentially at high risk of both VTE and bleeding complications. The recently approved lower prophylactic dose of fondaparinux appears to be a safe and relatively effective strategy in these patients.


Patients admitted to hospital for an acute medical illness, including heart failure, respiratory failure, infection and neurologic disease, are at increased risk of venous thromboembolism (VTE). Large randomized controlled trials (RCTs) have shown that pharmacologic prophylaxis with anticoagulant drugs effectively reduces this risk [1–3], and guidelines recommend the routine use of thromboprophylaxis in hospitalized acutely ill medical patients with risk factors for VTE [4]. Recommended strategies include low-molecular-weight heparin (LMWH), unfractionated heparin (UFH) or fondaparinux [4].

There are reports that about 40% of hospitalized medical patients have a moderate or severe renal insufficiency [5,6]. The presence of renal insufficiency is associated with both an increased risk of VTE and an increased risk of bleeding complications [7–10], especially because patients with kidney failure tend to be older and more frequently have co-morbidities than patients with normal renal function. Furthermore, the use of anticoagulant drugs with predominant renal excretion such as LMWH or fondaparinux may result in an excessive anticoagulant effect due to drug bioaccumulation, thus predisposing to bleeding. In 2008, the American College of Chest Physicians’ guidelines recommended that in the presence of renal impairment, in particular in elderly patients and in patients at high risk of bleeding, physicians should avoid the use of an agent that bioaccumulates or should consider using a lower dose of the agent or monitoring the drug level or its anticoagulant effect [11]. However, clinical information on the optimal use of anticoagulants in renally impaired patients is limited.

A lower dose of fondaparinux (1.5 mg daily) was recently approved for the prevention of VTE in these patients after the results of pharmacokinetic simulations based on data obtained in patients undergoing major orthopedic surgery showed a predicted exposure to the drug in patients with renal impairment similar to that observed with 2.5 mg in patients with normal renal function [12]. In order to explore the safety and efficacy profile of fondaparinux 1.5 mg in acutely ill medical patients with a creatinine clearance between 20 and 50 mL min−1, a prospective cohort study was carried out.


Study population

Inclusion criteria were the same as those used in the ARTEMIS study [3], a study that compared fondaparinux 2.5 mg once daily with placebo in older medical patients. Thus, acutely ill patients admitted to a medical ward, aged 60 years or older, and expected to remain in bed for at least 4 days, were eligible if they had at least one of the following: congestive heart failure class III or IV according to the New York Heart Association (NYHA) classification, acute respiratory failure in the presence of chronic lung disease, or clinically diagnosed acute infections or inflammatory disorders such as arthritis, connective tissue diseases or inflammatory bowel disease. In addition, patients were eligible if their creatinine clearance calculated on admission was between 20 and 50 mL min−1 based on the Modification of Diet in Renal Disease (MDRD) formula [13]. Written informed consent was obtained from the patients before inclusion in the study. Patients were excluded: if they had active bleeding or if they had bleeding events in the previous 3 months; in the case of known bleeding diathesis; if platelet count was < 100 000 per cubic milliliter; if they were receiving therapeutic doses of UFH, LMWH or fondaparinux; if they were receiving oral anticoagulant treatment with a vitamin K antagonist; if they received prophylactic doses of UFH, LMWH or fondaparinux for more than 48 h before inclusion in the study; or if their life expectancy was considered to be < 1 month.

Study design

This was a multicenter, investigator-initiated, prospective cohort study carried out in 20 Departments of Internal Medicine in Italy. The list of departments is provided in Appendix 1. The study was approved by the local ethics committees. After providing written informed consent, all patients received treatment with fondaparinux sodium (Arixtra®; GlaxoSmithKline, Verona, Italy) 1.5 mg daily s.c. for a minimum of 6 days to a maximum of 15 days. The first dose was administered in the evening of the day of inclusion in the study. Patients were followed-up for 1 month after discharge from the hospital. Therapeutic decisions after the end of the study treatment period were left to the discretion of the attending clinicians.

Study outcomes

The primary endpoint of the study was the incidence of major bleeding during treatment and 2 days thereafter. Secondary endpoints were the incidence of clinically relevant non-major bleeding during treatment and 2 days thereafter and the incidence of symptomatic VTE during treatment.

Major bleeding was defined as fatal bleeding, bleeding leading to surgical intervention, bleeding occurring in a critical organ (intracranial or intraspinal, retroperitoneal, intraocular resulting in visual impairment), or overt bleeding associated with a drop in hemoglobin levels of two or more grams/deciliter or requiring the transfusion of two or more units of red blood cells. Clinically relevant non-major bleeding was defined as any overt bleeding requiring a medical intervention and/or treatment discontinuation and not meeting any of the criteria for major bleeding. Symptomatic deep vein thrombosis was confirmed by compression ultrasound or computed tomography-scan; symptomatic pulmonary embolism was confirmed by high probability lung scan, pulmonary angiography or helical computed tomography (CT) scan. Death was classified as due to pulmonary embolism, bleeding or other causes. All suspected outcome events were adjudicated by an independent adjudication committee.

Study oversight

The study was designed by the Steering Committee members. Data were collected, maintained and analyzed by the Research Center on Thromboembolic Diseases and Antithrombotic Drugs of the University of Insubria in Varese, Italy. The adjudication committee independently assessed all outcome events by reviewing written reports from participating centers and, if necessary, clinical charts. The study was performed in accordance with the Declaration of Helsinki and local regulations. The protocol was approved by the Institutional Review Board or Ethic Committee at each study center. A written informed consent was obtained from all patients.

Sample size and statistical analysis

The pooled incidence of major bleeding events in the groups of patients receiving active thromboprophylaxis in the MEDENOX, PREVENT and ARTEMIS studies [1–3] was 0.6%, with an upper limit of the 95% confidence interval (CI) of 3.59%. To prove the safety of fondaparinux 1.5 mg in patients with moderate to severe renal insufficiency, we hypothesized that the upper limit of the 95% CI of the incidence of major bleeding events in this study would not have exceeded the cut-off of 3.59%. To prove this hypothesis, we calculated a sample size of 300 patients. However, because the enrollment rate was much slower than anticipated, recruitment was stopped after two and a half years with the inclusion of about 200 patients. All analyses were performed after the decision to interrupt the study.

Continuous variables were expressed as mean ± SD or as median with minimum and maximum values when data did not have a normal distribution (according to the Wilk-Shapiro test); categorical data are given as counts and percentages. Incidence of major bleeding, clinically relevant non-major bleeding and thromboembolic events was expressed as a percentage (with 95% CI with continuity corrections).

Baseline characteristics and prevalence of potential risk factors were initially compared in patients with and without major and clinically relevant non-major bleeding events. We then planned to construct a multivariate model introducing only variables with a P < 0.1 in the univariate model. Analysis was based on the principle of intention to treat. A P value of < 0.05 was chosen as the cut-off for statistical significance. The data were analyzed with the use spss software, version 19.0 (SPSS, Bologna, Italy).


Baseline characteristics

Between June 2009 and December 2011, a total of 206 patients were enrolled in the study. Baseline characteristics of the patients are summarized in Table 1. The mean age of the study population was 82 years, mean creatinine clearance was 33 mL min−1, two-thirds of patients were receiving concomitant antiplatelet agents, in most cases aspirin, and the mean Charlson co-morbidity index (see Appendix 2 for definition) was 8, with nearly all patients having an index of ≥ 5.

Table 1.  Baseline characteristics
  1. *The body weight was available in 164 patients (79.6%).

Patients, n206
Age (years), mean ± SD82.1 ± 7.1
Male, n (%)105 (51.0)
Weight (kg), mean ± SD*72.3 ± 16.1
Creatinine (mg dL−1), mean ± SD1.8 ± 0.5
Creatinine clearance (mL min−1), mean ± SD33.2 ± 8.2
Inclusion criteria, n (%)
 Congestive heart failure101 (49.0)
 Acute respiratory failure88 (42.7)
 Acute infection90 (43.7)
 Acute inflammatory disease12 (5.8)
 Inflammatory bowel disease2 (1.0)
Concomitant diseases, n (%)
 Diabetes mellitus74 (35.9)
 Hypertension163 (79.1)
 Obesity45 (21.8)
 Coronary artery disease69 (33.5)
 Previous stroke/TIA33 (16.0)
 Cancer45 (21.8)
 Previous major bleeding6 (2.9)
Concomitant treatment, n (%)
 Antiplatelets128 (62.1)
 NSAID9 (4.4)
Charlson score, mean ± SD8.2 ± 1.50
Charlson score ≥ 5, n (%)205 (99.5)
Duration of treatment with fondaparinux (days), mean ± SD9.2 ± 3.6
Delay between admission to hospital and the first dose of fondaparinux (days), median (range)1 (0–6)
Length of stay (days), median (range)11 (2–58)
Median duration of observation for the primary endpoint major bleeding (days, range)11 (1–17)
Duration of follow up (days), mean ± SD40.3 ± 11.5


One patient experienced major bleeding for a frequency of 0.49% (95% CI 0.03–3.10%; Table 2). This patient had gastrointestinal bleeding and required blood transfusions. There were eight non-major clinically relevant bleeding events (four episodes of macroscopic hematuria after catheterization, one rectal bleeding in a patient with internal hemorrhoids, one mouth bleeding in a patient with varices of the tongue, one epistaxis, and one gastrointestinal bleeding) for a frequency of 3.88% (95% CI 1.81–7.78%). One additional major bleeding event occurred during follow-up, at day 19, in a patient treated with fondaparinux for 14 days. Five days after hospital discharge the patient was readmitted because of relapsing pneumonia and severe anemia. Gastroscopy revealed the presence of duodenal ulcers with recent signs of bleeding.

Table 2.  Study outcomes
Outcomes n events/N% (95% CI)
Major bleeding1/2060.49 (0.03–3.10)
Non-major clinically relevant bleeding8/2063.88 (1.81–7.78)
Major bleeding + non-major clinically relevant bleeding9/2064.37 (2.15–8.40)
Symptomatic venous thromboembolism3/2061.46 (0.38–4.55)
Death23/20611.17 (7.36–16.48)

When we compared baseline characteristics of patients with and without bleeding events during treatment with fondaparinux, none of the assessed risk factors was significantly associated with bleeding at the univariate analysis (Table 3). Only age was of borderline statistical significance (P = 0.06). Given the results of univariate analysis, multivariate analysis was not performed.

Table 3.  Comparison of characteristics of patients with bleeding events (including major plus non-major clinically relevant bleeding) and patients without bleeding events
CharacteristicsBleeding (n = 9)Not bleeding (n = 197) P value
Mean age (years)
Age ≥ 80 years, n (%)8 (88.8)131 (66.5)0.28
Males, n (%)6 (66.6)99 (50.3)0.50
Mean creatinine clearance (mL min−1)32.733.30.82
Creatinine clearance < 30 mL min−1, n (%)3 (33.3)74 (37.6)1
Previous major bleeding, n (%)1 (11.1)5 (2.54)0.23
Concomitant antiplatelet therapy, n (%)5 (55.5)123 (62.4)0.73
Mean duration of prophylaxis (days)
Mean Charlson score8.68.20.44

Venous thromboembolism

Three patients experienced symptomatic VTE, which in all cases was pulmonary emboli with or without deep vein thrombosis; the frequency was 1.46% (0.38%–4.55%; Table 2). The first event was diagnosed 1 day after inclusion in the study and after the administration of a single dose of fondaparinux. The patient had chronic obstructive pulmonary disease and was admitted for severe dyspnea. Compression ultrasound revealed proximal deep vein thrombosis of the right lower limb and lung scan showed high probability of pulmonary embolism. The second event was an isolated pulmonary embolism diagnosed at helical CT scan after 13 days from hospital admission in a patient admitted for concomitant congestive heart failure and pneumonia. The third event was a pulmonary embolism with concomitant right lower limb deep vein thrombosis diagnosed by means of helical CT scan and compression ultrasound 5 days after hospital admission in a patient admitted for pneumonia. All events were treated with LMWH and warfarin and none of the events was fatal. No further symptomatic VTE events were reported during follow-up.


The incidence of mortality was 11.17%, and only in one case did the investigators consider pulmonary embolism as a possible cause, although an alternative diagnosis (fatal arrhythmia in a patient with known cardiac disease) was present. Ten deaths occurred while patients were on prophylaxis, 13 during follow-up. Worsening pneumonia (n, 8), decompensated heart failure (n, 4), metastatic cancer (n, 4) and septic shock (n, 3) were the most common causes of mortality.


This study was carried out in a population of medical patients with traditional risk factors for VTE and with moderate to severe renal impairment. Using this additional inclusion criterion, we identified a very old and truly fragile population, that is increasingly encountered in ‘real world’ clinical practice, but has been poorly represented in previous RCTs. This population is challenging, because the risk of bleeding associated with the use of pharmacologic thromboprophylaxis can outweigh the benefit from effective VTE prevention. In this study, the use of a lower dose of fondaparinux (1.5 mg) was associated with an acceptable rate of bleeding events, suggesting this approach can be safe in this setting.

The incidence of major bleeding events, 0.49%, was similar to the 0.6% overall incidence of major bleeding events calculated by pooling the results of three major RCTs in this setting, MEDENOX, PREVENT and ARTEMIS [1–3], and the upper limit of the 95% CI observed in this study (3.10%) remained lower than that calculated by pooling the results of the three trials (3.59%). In a meta-analysis of RCTs comparing anticoagulant prophylaxis with no treatment, the incidence of major bleeding was 0.58% in the group of treated patients [14]. The incidence of major bleeding observed in our study also compares favourably with that reported (1.2%) in a large observational study carried out in an unselected population, the IMPROVE study [9]. In IMPROVE, major predictors of bleeding were, among others, age > 85 years and severe renal insufficiency. In our study, 37.8% of patients were 85 years or older and 37.2% of patients had a creatinine clearance below 30 mL min−1. Furthermore, the majority of patients enrolled in our study were concomitantly receiving an antiplatelet agent, an additional risk factor for bleeding.

The incidence of symptomatic VTE in this study, 1.46%, was higher than that observed in a meta-analysis of RCTs in patients receiving active treatment (0.20% pulmonary embolism and 0.38% deep vein thrombosis)[14], but similar to that reported in observational studies carried out in unselected populations [5,15]. This difference is likely to be explained by the higher baseline risk for VTE in our population as compared with the populations enrolled in RCTs, and by the fact that in RCTs all patients undergo routine screening for DVT, thus possibly interfering with the natural clinical history of the disease and contributing to the reduction in the incidence of symptomatic events. Given the incidence of VTE in this study, a question may arise on whether a higher dose of fondaparinux (i.e. 2.5 mg) would have been more effective and would have been able to reduce this rate. However, given the fragility of this population and the concomitant need to minimize the risk of bleeding in order to offer an acceptable net clinical benefit, this theoretically supports our choice to reduce the doses of anticoagulant prophylaxis with fondaparinux in these patients. Indeed, only a head to head comparison between the higher and the lower dose of the drug could provide definite evidence on which approach is associated with the most favourable clinical benefit in this population.

Finally, 30-day mortality rates in this study were higher than in many of the previous RCTs carried out in this setting. In the five studies that assessed all-cause mortality in patients receiving prophylaxis identified by the meta-analysis of Dentali et al. [14], death occurred in 4.3% of patients. A similar rate (4.9%) was found in the recently published Lifenox study, a trial that aimed to assess whether enoxaparin could reduce mortality in medical patients as compared with placebo [16]. The higher fatality rate reported in our study is not surprising given the mean age of the population and, most of all, the severity of co-morbidities as defined by the Charlson score [17]. In our study, the mean score was 8.2 and all but one patient had a score of ≥ 5, a cut-off that identifies patients at the highest risk of mortality [17].

This study has a number of limitations. First, enrollment rate was slower than expected and the initially calculated target sample size was not reached. However, the confidence intervals of the estimated rate of major bleeding remained quite narrow, suggesting that the inclusion of an additional 94 patients may have not substantially modified our findings. Indeed, this difficult enrollment in particular at some centers may suggest the possibility of selection bias and may raise the question of generalizability of results. However, baseline characteristics of our patients describe a very old and severely ill population, and it is possible that many of the excluded patients were too sick to be good candidates for thromboprophylaxis. Unfortunately, a screening log was not collected and no information regarding excluded patients is available. Second, this was a single-arm cohort study, and the lack of a control group weakens the validity of our conclusions about the safety and efficacy of low-dose fondaparinux. Unfortunately, there is no widely accepted treatment regimen for patients with severe renal insufficiency and the identification of an adequate control group was therefore problematic. We believe that, based on our findings, low-dose fondaparinux could be an ideal control group for future studies in this setting.

Third, asymptomatic deep vein thrombosis was not routinely screened in this study, and some events have likely been missed. However, we focused on symptomatic events only because we were interested in major outcomes that directly affect the management of these patients; moreover, all patients were followed-up for a month after hospital discharge, thus reducing the possibility of missing important asymptomatic events that may have occurred prior to or during hospitalization and become symptomatic. Furthermore, the objective diagnosis of VTE and the central adjudication of study outcomes have also contributed to minimize the potential for bias. Finally, the optimal strategy to estimate renal function in clinical practice remains debated. The MDRD formula is standardized to a body surface area (BSA) of 1.73 m2 and it appears to be most accurate in patients with a BSA between 1.5 and 2.4 m2 [18]. In our study, of the 163 (79.1%) patients in whom BSA could be calculated (see Appendix 2), 91.4% were within this range, suggesting the MDRD formula was the appropriate method for estimating glomerular filtration rate in our population.

In conclusion, this study identified a population of acute medical patients at very high risk of both VTE and bleeding complications. The recently approved lower prophylactic dose of fondaparinux 1.5 mg once daily appears to be a safe and relatively effective strategy in these patients.


The steering committee reported in Appendix 1 (W. Ageno, F. Dentali, A. Squizzato, D. Imberti, M. Pini, M. Di Nisio, M. Silingardi, G. Agnelli) contributed to the conception and design of the study and interpretation of data. W. Ageno, N. Riva and F. Dentali contributed to acquisition and analysis of the data, and drafted the article. P. Noris, M. Di Nisio, M. La Regina, D. Arioli, L. Ria, V. Monzani, S. Cuppini, E. Lupia and M. Giorgi Pierfranceschi contributed to acquisition of the data, interpretation, and critical revision of the manuscript. All authors provided final approval of the manuscript. The principal investigator (W. Ageno) had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.


The study was supported by a grant in aid by GlaxoSmithKline, Verona, Italy. We are indebted with A. Cohen and A. Gallus for their critical review of this manuscript and their invaluable suggestions. identifier: NCT00927602.

Disclosure of Conflict of Interests

WA has sat on advisory boards for Bayer, BMS/Pfizer and Daiichi Sankyo, and has received honoraria for speaking activities for Boheringer Ingelheim, Bayer, BMS, Pfizer, Sanofi and GlaxoSmithkline. The other authors have no conflicts of interest to disclose.


Appendix 1

The members of the FONDAIR Study were as follows. Steering Committee: Walter Ageno, Francesco Dentali, Alessandro Squizzato, Davide Imberti, Mario Pini, Marcello Di Nisio, Mauro Silingardi and Giancarlo Agnelli. Writing Committee: Walter Ageno, Francesco Dentali and Nicoletta Riva. Study and Data Management: Raffaella Cavi and Nicoletta Riva. Central Independent Adjudication Committee for recurrence of venous thromboembolism and bleeding: Marco Moia (Chairman), Franco Piovella and Gualtiero Palareti.

FONDAIR investigators (at least one patient enrolled)

Department of Clinical Medicine, University of Insubria, Ospedale di Circolo, Varese (Luigi Steidl, Luca Galli, Stefano Grazioli, Sara Turato; 27 patients); IRCCS Policlinico San Matteo Foundation, University of Pavia, Pavia (Francesca Bellistri; 26 patients); Department of Internal Medicine, University of Chieti, Ospedale Annunziata, Chieti (Ettore Porreca; 24 patients); Department of Internal Medicine, Ospedale S. Andrea, La Spezia (Luigi Orlandini; 15 patients); Department of Internal Medicine, Arcispedale S. Maria Nuova, Reggio Emilia (Attilia Maria Pizzini; 14 patients); Department of Internal Medicine, Ospedale Sacro Cuore di Gesù, Gallipoli (Luigi Ria; 13 patients); Department of Emergency Medicine, IRCCS Ospedale Maggiore Policlinico, Milano (Silvia Serafini; 12 patients); Department of Internal Medicine, Ospedale di Rovigo, Rovigo (Ciro Rossetti, Marco Marzolo, Emilio Ramazzina; 12 patients); Department of Internal Medicine, University of Turin, Ospedale S. Giovanni, Torino (Giuseppe Montrucchio; 11 patients); Department of Internal Medicine, Ospedale di Piacenza, Piacenza (10 patients); Department of Internal Medicine, Ospedale S. Biagio, Domodossola (Egidio De Gaudenzi; 9 patients); Department of Internal Medicine II and Emergency Medicine, Ospedale della Carità, Novara (Roberta Re, Mauro Campanini; 8 patients); Department of Internal Medicine, Ospedale di Fidenza, Fidenza (Corrado Pattacini; 6 patients); Department of Internal Medicine, Ospedale del Delta, Ferrara (Roberto Manfredini, Emanuela Rizzioli; 5 patients); Department of Clinical Medicine, University Hospital of Sassari, Sassari (Gianpaolo Vidili, Maria Luisa Cappai, Giuseppe Delitala; 3 patients); Department of Internal Medicine, Ospedale Vito Fazzi, Lecce (Fernando Parente; 3 patients); Department of Internal and Vascular Medicine, University of Perugia, Ospedale S. Maria della Misericordia, Perugia (Cecilia Becattini; 3 patients); Department of Internal Medicine, Ospedale C. Magati, Scandiano, Reggio Emilia (Giuseppe Chesi, Patrizia Zoboli; 3 patients); Department of Internal Medicine, University of Rome, Policlinico Umberto I, Roma (Francesco Violi, Stefania Basili, Lorenzo Loffredo); Department of Internal Medicine, IRCCS Ospedale Maggiore Policlinico, Milano (Guido Moreo, Denis Fiore; 1 patient).

Appendix 2

Charlson weighted index of co-morbidity (17).

Assigned weight for diseasesConditions
1Myocardial infarction
Congestive heart failure
Peripheral vascular disease
Cerebrovascular disease
Chronic pulmonary disease
Connective tissue disease
Ulcer disease
Mild liver disease
Moderate or severe renal disease
Diabetes with end organ damage
Any tumor
3Moderate or severe liver disease
6Metastatic solid tumor

Body surface area (BSA) was calculated using Mosteller formula: BSA (m2): ([Height(cm) × Weight(kg)]/3600)1/2).