Updates on the Risk of Dalteparin Thromboprophylaxis Failure in Hospitalized Acutely Ill Medical Patients with Cancer (ARTIC STUDY)

Background

Hospitalization is a significant risk factor for venous thromboembolism (VTE) with  25% of all VTE  occurring in this setting. In patients with cancer this risk may be higher due to the inherent procoagulant state malignancy-induced, -cancer therapy and its complications. Current  oncology guidelines suggest that hospitalized medical patients receive venous thromboprophylaxis with unfractionated heparin or low molecular weight heparin if their hospital stay is longer than 3 days.

We sought to evaluate if patients with cancer hospitalized for management of an acute medical illness are at higher risk for failing standard anticoagulation prophylaxis with dalteparin compared to historical data.

Methods

This is a single- centre retrospective cohort study (London, Canada). We included adult patients; with any type of active cancer admitted for at least 3 days for treatment of an acute medical reason, who received prophylaxis with dalteparin during  hospital stay.

Acute medical illness was:  failure to thrive; fever; need for cancer treatment as an inpatient;  pain control;  acute respiratory illness. The main study outcome was failure of VTE prophylaxis defined as symptomatic and objectively diagnosed  pulmonary embolism (PE) or deep venous thrombosis (DVT)  within 3 months of the most recent hospital discharge.

Results

Between January 2011 and December 2013 our hospital registered 4262 admissions of patients with cancer for treatment of an acute medical illness. 875 patients (1132 admissions) fulfilled our eligibility criteria.   681 (78%) patients were classified as having a single admission. Of those, 247 had  previous but excluded admissions (“pseudo-single”), which leaves 434 patients with true single admissions. In total, there were 434 (49.5%) were males, mean age was 64.3 (SD= 13.5). Primary tumor sites were hematological (n=180); genitourinary (n= 170); lung (n=158); gastrointestinal (n=128) and others (n= 289).559 (70%) patients had stage III or IV. Reasons for admission were failure to thrive (n= 232; 26.6%); fever (n= 202; 23.3%); need for cancer treatment as  inpatient (n= 154; 17.7%); pain  (n=126; 14.5%); respiratory distress(n= 108; 11.6%) or pain (n= 53; 6.3%). Mean hospitalization days was 14.7 (±12). 

VTE occurred in 78 of 875 (8.9%) patients or 78 of 1132 admission (6.9%): 36(46%)  DVT, 34 (43.5%) PE and 7 PE + DVT (8.9%). 34 of 78 (43.6%)  VTE occurred within the first 14 days of admission. However, the overall risk for VTE appeared to be much more significant if the patient remained hospitalized for more than 14 days [RR=3.7 (95%CI= 1.99 – 4.71; p<0.001)]. . 150 (15%) patients with single admissions had a concomitant diagnosis of cancer and VTE (within 30 days of admission).

A univariate analysis suggested that having had multiple admissions (OR=0.3; 95%CI=0.17- 0.54; p=0.008); being man (OR= 1.69; 95%CI: 1.03 – 2.78; p=0.039); being admitted due to respiratory distress (OR=2.6; 95%CI: 0.9 – 6.8; p=0.052) or failure to thrive (2.52; 95%CI: 1.06 – 5.9; p=0.036) were potentially significant predictors of VTE risk. However, the logistic regression confirmed that the only significant risk factor is the number of admissions (Table).

When we compared our results with the pooled data from the MEDENOX, PREVENT and ARTEMIS trials we found that the incidence of VTE in the cancer patients included in the trials was the same as ours [13 of 143 patients (8.8%)] and the incidence of VTE in those without cancer was 95 of 2139 (4.5%), significantly lower compared to the cancer cohort (p=0.0004).

Conclusion

Our study suggests that hospitalized patients with active cancer are at high risk for VTE prophylaxis failure (8.9%) and our results are in keeping with the  literature. It appears that the most important risk factor for thromboprophylaxis failure is having a first admission as a cancer patient. New VTE prophylactic strategies for this population should be investigated in future prospective studies.

Table- Logistic Regression of potential predictors of VTE risk in hospitalized cancer patients

Variable	Odds Ratio	95% CI		p-value
Male	1.05	0.64	- 1.74	0.8270
“Pseudo-single” admission*	0.53	0.28	- 0.97	0.9182
Multiple admission*	0.29	0.16	- 0.53	0.0029
Respiratory distress	1.19	0.53	- 2.68	0.7597
Pain	1.27	0.59	- 2.71	0.5079
Failure to thrive	1.02	0.46	- 2.24	0.6820

*Reference: single admission