Venous thromboembolism (VTE) is a common complication of neurosurgery, but since this condition often has no symptoms, identifying patients with VTE can be challenging. Many neurosurgeons view one diagnostic tool, the D-dimer test, skeptically because D-dimer, a product of fibrin clot degradation, rises naturally from surgery, making it hard to interpret a high D-dimer test result postoperatively. Researchers recently hypothesized that the increase in D-dimer levels associated with craniotomy would be systematic and predictable, whereas craniotomy patients with VTE would have significantly higher levels. Their findings are the subject of this issue of Strategies.
D-dimer levels rise not only in the presence of VTE, but also a variety of other conditions, including cancer, stroke, infection or renal disease, and most particularly, surgery. Distinguishing normal increases caused by activation of the coagulation cascade in response to surgery versus a pathological clot remains challenging. Yet neurosurgeons long for a better means of doing so to facilitate rapid treatment of and better patient outcomes associated with VTE.
“Venous thromboembolism is quite commonplace in neurosurgery. No one who has been in neurosurgery a few years does not remember a case of a perfectly healthy patient coming out of surgery with no neurologic deficit suddenly being struck down by venous thromboembolism and simply dying. So this is something neurosurgeons are really interested in, to find a feasible screening method for venous thromboembolism, as there are counter measures for it,” explained the study’s lead author, Julian Prell, MD, an attending neurosurgeon at the University of Halle in Halle, Germany.
Prell went on to explain his interest in taking a closer look at D-dimer. “For many years people have said D-dimer doesn’t work with patients who’ve been operated on recently because blood coagulation is activated by surgery and you won’t have values you can believe. This was the accepted hypothesis throughout neurosurgery,” he said. “However, it has never been demonstrated that D-dimer results are not systematically raised by surgery. So we conducted this study with the hypothesis that D-dimer would be raised postoperatively but would be raised even more if the patient actually experienced a venous thromboembolism.”
Prell and his colleagues conducted a prospective study of 101 patients who underwent elective craniotomy (J Neurosurg 2013; doi: 10.3171/2013.5.JNS13151). In addition to pre- and postoperative Doppler ultrasonography, the study protocol involved D-dimer testing preoperatively and on days 3, 7, and 10 postoperatively. Overall, 43% of patients experienced VTE postoperatively.
The researchers found non-significant differences in mean preoperative D-dimer levels between patients who went on to develop VTE in comparison to those who didn’t, 0.74 mg/L versus 0.62 mg/L, respectively. Postoperatively, however, the investigators observed marked differences in D-dimer values between patients who developed sonography-confirmed VTE. This difference was greatest on the third day, when patients who developed VTE had a mean D-dimer value of 5.49 mg/L versus 1.59 mg/L in those who did not have VTE. The differences, although smaller, persisted on days 7 and 10 and remained statistically significant.
The researchers considered five possible cut-off values based on optimum sensitivity for VTE (0.86 mg/L); optimum specificity for VTE (4.3 mg/L); optimum balanced sensitivity and specificity (3.3 mg/L); optimum clinical applicability (2 mg/L); and the manufacturer-recommended cut-off for the Siemens Innovance D-dimer test (0.5 mg/L). These different cut-offs yielded a wide range of positive- (PPV) and negative-predictive values (NPV). For example, the cut-off aimed at optimum specificity (4.3 mg/L) resulted in a 100% PPV and 75% NPV, while the one designed to optimally balance sensitivity and specificity (3.3 mg/L) yielded 89.5% PPV and 85.7% NPV. The authors’ proposed clinically optimal cut-off (2 mg/L) captured all but two patients with VTE, yielding a PPV of 73.2% and NPV of 95.6%.
“While not perfect, this test appears to be significantly better than clinical suspicion as the sole ‘screening’ method,” the authors wrote. “In our study group, 8.9% of all patients presented with clinically apparent deep vein thrombosis [DVT] …the general incidence of DVT was more than four times higher than the incidence of symptomatic DVT. Given these proportions, more than three of four patients with DVT would remain undiagnosed and at risk without D-dimer screening.”
Prell was quick to point out the limitations of the study in that it involved only elective craniotomy patients from one center. Still, he and his authors believe their findings to be transferrable in principle to other homogenous patient populations, and he called on colleagues elsewhere to conduct similar experiments. “There might be indications for which the D-dimer test can’t be used because the activation of the coagulation system by the surgery is so immense that it really will mask the activation by the thrombosis. We simply don’t know yet,” said Prell. “If I were a thoracic or orthopedic surgeon I would give this a try. It’s not a very complicated study to run and might provide interesting results.”
A clinician-researcher not involved in the study also called for further investigation of this issue. “I think their approach is a very sensible one, but I would consider this study to be hypothesis-generating. What they need to do now is prospectively test that hypothesis by doing a second study and show that this cut-point is indeed valid,” suggested Jeffrey Ginsberg, MD, FRCPC, professor of medicine in the department of hematology and thromboembolism at McMaster University in Hamilton, Ontario, Canada. “We’d like to see other studies replicate the results. That would be very strong evidence that their hypothesis-generating study is a real finding, not a chance observation.”
Ginsberg went on to explain that similar studies he has been involved in and others that have not been published did not identify a clear discriminatory point that differentiated normal from abnormal patients. Prell and his colleagues indeed are conducting another study involving D-dimer testing on the third day after craniotomy, followed by—depending on the results—Doppler sonography. They also plan to conduct a similar study involving spine surgery