Cloud Hosting

We evaluated the ability of the PPS to predict VTE in non-surgical hospitalized patients. The results did not demonstrate significant benefit of the PPS or the prophylactic anticoagulation based on it. The occurrence of recorded clinically significant VTE events in non-surgical patients during hospitalization and up to 90days thereafter was relatively low, about 1 event in 400 admissions, compared to previously reported estimations5,11.

The study results demonstrated a difference of 0.01% between the risk groups (0.27% vs. 0.28%) and the statistical analysis indicated that this difference in the studied sample size of about 15,000 patients is not statistically significant. The calculated sample size to show statistical significance of a difference of 0.1% (tenfold higher than the observed difference) is about 80,000 patients and for a difference of 0.01% (as was observed in the study) it is over 8 million patients. This means that a profoundly high number of patients would have to be treated to prevent a single VTE case, which exclude any clinical significance. The potential harm and high cost dramatically outweigh any benefit in this case.

The use of PPS was also not related to improved or worsen safety regarding bleeding events. The routine performance of PPS and the implementation of pharmacologic prophylaxis following PPS estimations involve substantial resources (medical personnel time, medication cost) and may distract physicians. Lack of beneficial effects calls to consider discontinuation of this practice to save resources.

While the presented study followed the Padua study outline, there is a marked difference between the VTE incidences observed in the studies, over tenfold. The diagnosis of VTE events in the Padua study was mainly based on positive imaging and laboratory (d-dimer) results and not on clinical symptoms5. A high percentage of the Padua study participants was referred to testing. Data regarding the clinical significance of the VTE diagnosis in the Padua study are limited. Contrary, the presented study identified recorded symptomatic clinically meaningful VTE events that demanded intervention or readmission. The different methods may have led to the different outcome rates. In addition, the duration of hospitalization was relatively longer in the Padua study, which may have contributed to the elevated VTE prevalence.

Others also found low efficiency of the PPS. Saliba et al. measured single thrombin generation in acutely hospitalized patients and found no correlation with the PPS12. Vardi et al. showed that PPS estimation lacks granularity in detecting non-surgical septic patients at risk of acquiring VTE13. Depietri et al. assessed the application of risk assessment models, including the PPS, on VTE and major hemorrhage on internal medicine hospitalized patients and found it statistically ineffective14.

The PPS was compared to other risk assessment tools. The Caprini risk assessment model (RAM) performed better than the PPS in 2 studies. Zhou et al. assessed the validity of the Caprini RAM in Chinese hospitalized patients with VTE and compared it to the Kucher tool and the PPS15. Caprini model was found to classify much more VTE patients into high or highest risk level than the other models with statistically significant differences (Caprini model vs Kucher model, p<0.0001; Caprini model vs the PPS, p<0.0001). In another study by Zhou et al. the Caprini RAM was compared again to thePPS16. The Cparini RAM defined 82.3% of VTE cases as high risk, while the PPS has defined only 30.1% of these same cases. Nendaz et al. investigated the Geneva risk score among Swiss hospitalized medical patients17. They found that for VTE prediction, the Geneva Risk Score compared favourably with the PPS, particularly for its accuracy to identify low-risk patients who do not require thromboprophylaxis. The main practical limitation of the Caprini and Geneva models is their relative complexity, as they include 39 and 19 criteria, respectively, compared to the 11 criteria of PPS.

PPS is a mathematical scoring method based on the models of Kucher et al. and Lecumberri et al. that uses some of the criteria in the original models18,19. Its ability to capture the full clinical complexity of different patients and become a globally generalized valid assessment tool is limited. Risk assessment should be comprehensive and appropriately tailored to the patients and the relevant population. In addition, risk should be continuously evaluated and not assessed only at a single time point, such as admission.

The issue of any VTE diagnosis vs. symptomatic VTE is paramount. The practical contribution of asymptomatic VTE diagnosis to patients health and prognosis can be debated. Hospital related VTE was considered a safety concern following 2 studies published in the late 1980s20,21. Both were based on autopsies of patients that died during hospitalization or shortly after. These studies only demonstrated the presence of VTE in the body; there was no proof of causality or any linkage between VTE post-mortem presence and either morbidity or mortality. The Worchester study from 1991 showed higher VTE rates in hospitalized patients compared to community; this study also reported mainly asymptomatic cases22. About 10years later (20022003), 3 large-scale clinical trials sponsored by pharmaceutical companies were published almost simultanesly23,24,25. They all demonstrated significant improvement in VTE incidence following prophylactic anticoagulants. A critical review of the studies by Vardi and Haran emphasized that most of the VTE cases in the clinical trials were asymptomatic26. Symptomatic VTE rates were low and there was no benefit for the prophylaxis when observing only symptomatic cases.

During the years, there were several other publications that supported the importance of VTE risk during hospitalization and immediately after but based their conclusion on previous papers that mainly cited the old autopsy studies. Pradonicoauthor of the Padua studycited Leizorovicz and Mismetti to support in his paper the claim regarding VTE elevated risk in non-surgical hospitalized patients27,28. Leizorovicz and Mismetti themselves have based this claim on the THRIFT study paper29, which provides as references the 1980s autopsy studies, again20,21. The claim that the risk for clinically meaningful VTE in non-surgical hospitalized patients is significantly elevated, is poorly substantiated on clear clinical data.

Diagnosis of asymptomatic VTE is increasing with the advances in diagnostic technologies. Computed tomography pulmonary angiography is able to pick up pulmonary emboli as small as 23mm in diameter, but their clinical relevance is questioned30,31. Unnecessary anticoagulation increases the risk for major bleedings32.

The pathophysiology of VTE is related to an imbalance of 3 components of the Virchovs triangle: hypercoagulability, stasis, and endothelial impairment33. This imbalance does not necessarily worsen upon admission to hospitalization for most non-surgical patients; it can even improve. The risk for VTE is inherent and relatively fixed due to the patients prolonged personal medical and functional status before admission. Many of these patients are functionally impaired and immobile long before their hospitalization, while immobility of hospitalized patients is usually properly addressed by repeated position changes and other means and procedures, as instructed by current guidelines and hospital accreditation standards34. Blood viscosity is improved upon admission of deteriorated or dehydrated patients using fluid therapy. Infection and inflammation are appropriately and rapidly managed in most admitted patients. These interventions during hospitalization, which are nowadays a common practice, reduce the risk for VTE and the need for pharmacological prophylaxis. Shortening of the hospitalization is another key element of mitigating the risk.

Recent studies have indicated a low incidence of symptomatic VTE in hospitalized non-surgical patients. Fritz et al. found that only 0.3% of general medicine hospitalized patients developed VTE35. Koren et al. evaluated over 500 Israeli medical hospitalized patients and did not find any hospital acquired symptomatic VTE36. Vardi et al. studied septic patients in Israeli internal medicine departments and found a low rate of symptomatic VTE13. Kolomansky et al. evaluated prospectively the occurrence of VTE in Israeli hospitalized patients. Overall, VTE was diagnosed in 0.25% of these patients37. These reports support the presented study results.

Another support to the low incidence of clinically significant VTE in hospitalized patients comes from a recently published OECD report on VTE rates38. The recorded multinational multicenter VTE hospital-related diagnosis is 0.3% (189.3 DVT cases and 175.3 PE cases of 100,000 discharges), similar to the present study results. It represents the true extent of the clinically meaningful VTE cases.

The study has limitations. It is retrospective and based of computerized retrieval of digital data. It is depended on the heterogenic quality of medical information recorded by numerous staff members. It is a single center study and the generalization of its conclusion should be done carefully.

Nevertheless, the study holds several strengths. It has a large patients sample, over 5,000. More than twice the sample of the Padua study, which was also a single center study. This study outline replicated the Padua study group design and patient selection to enable a fair chance for comparison. The study period is relatively long, over 2years, and the outcomes were measured up to 90days after discharge. The data were pooled from 2 medical databases providing comprehensive information from the hospital and community clinics; this gives high credence to the results. An independent focused validation was performed to ensure the appropriateness of the VTE diagnosis.

In conclusion, PPS was not found to be an efficient tool for identification of non-surgical hospitalized patients with high risk for clinically significant VTE events. Prophylactic anticoagulation based on PPS did not provide significant clinical benefit. The study results indicate the need to consider stopping the use of PPS as a mandatory or preferred assessment tool for VTE in Israeli non-surgical hospitalized patients. Until confirmed, better and efficient assessment tools can be implemented, it is recommended to clinically assess VTE risk and the need for prophylaxis in hospitalized non-surgical patients on a case by case basis.

Read more:

Evaluation of the Padua Prediction Score ability to predict venous thromboembolism in Israeli non-surgical hospitalized patients using electronic...