Background: A variety of risk prediction scores have been devised to identify patients at increased risk for VTE in different patient populations and settings. Guideline recommendations for VTE risk assessment vary greatly. We performed a systematic review to identify and synthesize evidence on clinical risk prediction scores for VTE in medical and surgical hospitalized patients.
Methods: We systematically searched Medline, EMBASE Cochrane, NICE, NGC, and GIN databases up to March 2016. We included validation studies examining risk prediction scores for adult hospitalized patients. We excluded studies for any of the following reasons: non-English publication, conducted in non-OECD countries, validation cohorts focused solely on critical care patients, or validation of risk prediction scores developed for specific surgical or medical sub-specialty populations. We plotted ROC curves of included studies and performed summary ROC meta-analyses (DerSimonian-Laird random effects method, Open Meta Analyst software) for scores in which ≥1 external validation studies were combinable. Risk of bias was assessed qualitatively. We assessed the strength of the evidence base using the GRADE approach.
Results: No existing systematic reviews or meta-analyses were identified. We performed full text screening of 110 primary studies and included 18 of those for analysis. We included 6 validation studies of the Caprini score in surgical patients and 1 study in medical patients; 2 studies of the IMPROVE score in medical patients; 3 studies of the Padua score in medical patients; 1 study each of the Arcelus, Geneva and Kucher scores in medical patients; 1 study of the Khorana score in cancer patients; and 1 study of the RAP score in trauma patients. Strength of evidence was downgraded for study risk of bias because most studies disproportionately included patients at high risk of VTE. Our summary estimates of the performance of the 3 combinable scores at clinically-relevant thresholds are as follows: Caprini score at a threshold of 3 in surgical patients has 96% sensitivity and 44% specificity (Figure 1), IMPROVE at a threshold of 1 in medical patients has 96% sensitivity and 20% specificity (Figure 2), and Padua at a threshold of 4, the only reported threshold for this score in medical patients, has 87% sensitivity and 58% specificity.
Conclusions: There is moderate strength evidence for use of the Caprini score to predict VTE in surgical patients and for the Padua and IMPROVE scores in medical patients. Lower thresholds may be warranted to achieve sufficient sensitivity to identify low risk populations who may not require routine VTE prophylaxis. Studies making direct comparisons of risk prediction scores in similar patient populations are lacking and are necessary to ascertain which score is most effective.
