Clinical background
The Caprini score was constructed to structure the decision about pharmacological thromboprophylaxis in hospitalised surgical and medical patients. The decision is difficult because VTE risk varies widely between individuals and procedures, and prophylaxis itself carries a bleeding risk. Without a systematic risk assessment, prophylaxis tends either to be overused in low-risk patients or withheld from patients who would genuinely benefit. The main value of the score is that with a single summary measure it moves low-risk patients away from a presumed indication for prophylaxis, while flagging patients whose risk is so high that mechanical prophylaxis alone is insufficient.
Calculating the Caprini score
The Caprini score is a weighted sum of risk factors in which each factor is assigned 1, 2, 3 or 5 points according to its risk weight:
where according to age group and for each applicable risk factor. Age scores 0 points at years, 1 at 41 to 60 years, 2 at 61 to 74 years and 3 at years. One-point factors include, among others, a BMI > 25 kg/m², varicose veins, sepsis, serious lung disease, heart failure and medical bed rest. Two-point factors include general surgery > 45 min, malignancy, confinement to bed > 72 hours and a central venous catheter. Three-point factors comprise previous VTE, a family history of VTE and a number of thrombophilias, including factor V Leiden, the prothrombin 20210A mutation, lupus anticoagulant and heparin-induced thrombocytopenia. Five points are given for stroke within one month, elective major lower-limb arthroplasty, fracture of the hip, pelvis or leg, and acute spinal cord injury with paralysis.
The risk categories according to the 2005 revision are:
| Score | Category |
|---|---|
| 0–1 | Low |
| 2 | Moderate |
| 3–4 | High |
| ≥5 | Highest risk |
The model was originally developed in 1991 from data on 538 patients and revised in 2005 with an expanded list of risk factors and adjusted point weights [1,2]. The 2005 revision defined the four risk categories set out above. It is designed to predict composite VTE (deep vein thrombosis and pulmonary embolism) within 30 days of assessment [2].
Interpretation in practice
The risk categories translate into the choice of prophylaxis according to the following principles:
| Category | Score | Recommended action |
|---|---|---|
| Low | 0–1 | Early mobilisation; mechanical prophylaxis if needed |
| Moderate | 2 | Mechanical prophylaxis (intermittent pneumatic compression); pharmacological prophylaxis may be considered |
| High | 3–4 | Pharmacological prophylaxis (LMWH or low-dose unfractionated heparin) combined with mechanical prophylaxis |
| Highest risk | ≥5 | Pharmacological prophylaxis; consider extended duration after major cancer or orthopaedic surgery |
For patients with a score >8, studies have shown a VTE incidence of up to 11.3% without chemoprophylaxis within 60 days [3]. In this group, events occur not only immediately after surgery but are spread across the whole follow-up period, which is why extended prophylaxis for up to 30 days is recommended by several authors [3]. This is supported by observational data from both plastic surgical and general surgical cohorts [3].
For patients at highest risk after major cancer surgery or lower-limb orthopaedic procedures, the evidence for extended prophylaxis is strong. Randomised trials of LMWH for 7 versus 28 days after surgery for intra-abdominal or pelvic malignancy showed significantly fewer VTE events with the extended duration [3].
Validation and performance
The largest validation in surgical patients was reported by Bahl and co-workers in more than 8,000 patients in general surgery, urology and vascular surgery, with VTE incidences of 1.3% at a score of 5–6, 2.6% at 7–8 and 6.5% at >8, all significantly differentiated [3]. This led to a score of >8 being singled out as an extreme-risk group.
Pannucci and co-workers validated the 2005 version in the VTEPS network in 1,126 plastic surgery patients not receiving chemoprophylaxis at five US tertiary centres [3]. At 60 days the overall VTE incidence was 1.69%. A score of >8 was associated with a markedly increased risk compared with a score of 3–4 (OR 20.9, p < 0.001), 5–6 (OR 9.9, p < 0.001) and 7–8 (OR 4.6, p = 0.015). The incidence in the >8 group reached 11.3%, roughly twice as high as in Bahl's cohort, which is partly explained by the longer follow-up (60 versus 30 days) and by the fact that all VTEPS patients were without chemoprophylaxis.
A systematic review by Hayssen and co-workers examined 57 studies that had used the Caprini score to assign risk categories and reported the corresponding VTE rates [2]. The review demonstrated considerable heterogeneity in implementation: only 25% of studies used the validated cut-off levels, while 46% used four categories but with different thresholds. The VTE rate for the lowest risk category ranged from 0% to 12.3% and for the highest from 0% to 40%. Only 30% of studies measured the outcome at 30 days, the time point for which the model is intended [2].
A comparative study by Gibbs and co-workers at a US orthopaedic clinic tested Caprini against an institution-specific model in 80 patients undergoing elective hip or knee arthroplasty [4]. The mean Caprini score was 9.50 in patients with VTE and 9.35 in controls (p = 0.797), and the AUC was 0.52, corresponding to no discrimination at all. The reason is mechanical: the 2005 version assigns all patients undergoing elective lower-limb arthroplasty 5 points by default, so that all of them fall into the highest risk category regardless of other factors.
Limitations
The most important limitation concerns elective lower-limb orthopaedic surgery. Since the 2005 version gives 5 points for arthroplasty in itself, all such patients fall into the highest risk category. The score therefore cannot discriminate within this population, and the AUC has been measured at 0.52 [4]. A 2013 revision of the model subdivided the scoring and added factors such as smoking, blood transfusion and a BMI > 40, but the 2005 version lacks this differentiation [2,4].
The model has mainly been validated in surgical patients. Application to medical inpatients is less well studied, although some validations exist for medical patients on bed rest [2,3].
The Caprini score predicts composite VTE and does not distinguish the risk of deep vein thrombosis from the risk of pulmonary embolism [2]. It refers to events within 30 days, but many published validation studies use other follow-up periods, which makes direct comparison between studies difficult [2].
The systematic review showed that heterogeneity in implementation is so great that the same risk categories can correspond to VTE rates from 0% to over 40% depending on the study [2]. This underlines that the score should be interpreted in relation to the population and the procedure at hand, not as an absolute risk figure.
Finally, several factors may apply simultaneously, and some patients accumulate points from related but not independent factors, for example malignancy and major surgery for that same malignancy. This may lead to overestimation of the true increase in risk.
References
- Caprini JA. Thrombosis risk assessment as a guide to quality patient care. Dis Mon 2005;51(2-3):70-8. PMID: 15900257
- Hayssen H, Cires-Drouet R, Englum B et al. Systematic review of venous thromboembolism risk categories derived from Caprini scores. J Vasc Surg Venous Lymphat Disord 2022;10(6):1401-1409.e7. PMID: 35926802
- Pannucci CJ, Bailey SH, Dreszer G et al. Validation of the Caprini risk assessment model in plastic and reconstructive surgery patients. J Am Coll Surg 2011;212(1):105-112. PMID: 21093314
- Gibbs B, Paek S, Wojciechowski N et al. A comparison of the Caprini score with an institutional risk assessment tool for prediction of venous thromboembolism after total joint arthroplasty at an urban tertiary care health safety net hospital. Arthroplast Today 2023;23:101194. PMID: 37745953
