Table 6

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study weaknesses
NPV = Negative predictive value, NLR = Likelihood ratio for negative result.
Wells PS et al, 1995, Canada	214 consecutive patients referred for investigation of ?DVT.	Prospective cohort	Prevalence	25%	No sample size calculation Excluded patients with inconclusive venograms or plethysmogram
			Sensitivity	88% (CI 77, 96)
			Specificity	77% (CI 63, 80)
			NPV	95% (CI 89, 98)
			NLR	0.16
Brenner B et al, 1995, Israel	86 consecutive patients referred for investigation of ?DVT	Prospective cohort	Prevalence	58%	Small patient numbers. No sample size calculation. No confidence intervals
			Sensitivity	94%
			Specificity	61%
			NPV	88%
			NLR	0.1
Turkstra F et al, 1996, Netherlands	234 consecutive patients referred for ?DVT or ?PE	Prospective cohort	Prevalence	27%	No sample size calculation (but good numbers)
			Sensitivity	100% (CI 95, 100)
			Specificity	58% (CI 50, 65)
			NPV	100% (CI 96, 100)
Janssen MC et al, 1997, Netherlands	132 patients referred to ED or OPD for investigation of ?DVT	Prospective cohort	Prevalence	67%	No sample size calculation. Technique of assay may have affected results. Reference standard not applied to all patients
			Sensitivity	61% (CI 51, 71)
			Specificity	90% (CI 81, 99)
			NPV	52% (CI 29, 75)
			NLR	0.43
Ginsberg FS et al, 1997, Canada	398 consecutive patients referred to thromboembolic OPD as first episode of ?DVT	Prospective management study	NPV d-dimer alone	97.1% (CI 94.5, 98.8)	No sample size calculation. Reference standard not applied to all patients
			NPV d-dimer and plethysmography together	98.5% (CI 96.3, 99.6)
Mayer W et al, 1997, Austria	108 consecutive patients referred to vascular laboratory as ?DVT	Prospective cohort	Prevalence	31%	Small patient numbers. No sample size calculation. Used single ultrasound as reference standard
			Sensitivity	100% (CI 89, 100)
			Specificity	75% (CI 63, 84)
			NPV	100% (CI 94, 100)
Wildberger JE et al, 1998, Germany	250 consecutive patients referred for venography	Prospective cohort	Sensitivity	96%	No sample size calculation. Patient selection bias. No confidence intervals
			Specificity	59%
			NPV	97%
			NLR	0.06
Wells PS et al, 1998, Canada	496 consecutive outpatients referred with ?DVT	Prospective cohort	Overall sensitivity	94%	No sample size calculation. Patient selection bias. No confidence intervals
			Overall specificity	71%
			NPV	98% (CI 96, 99)
			NLR	0.08
			Low pretest probability
			Sensitivity	87%
			Specificity	76%
			NPV	99% (CI 97, 100)
			NLR	0.17
			Medium pretest probability
			Sensitivity	89%
			Specificity	64%
			NPV	97% (CI 90, 99)
			NLR	0.17
			High pretest probability
			Sensitivity	98%
			Specificity	54%
			NPV	86% (CI 42, 97)
			NLR	0.04
Mauron T et al, 1998, Switzerland	45 consecutive outpatients referred with ?DVT.	Prospective cohort	Prevalence	33%	Small patient numbers. No sample size calculation. Wide confidence intervals
			Sensitivity	53% (CI 28, 78)
			Specificity	70% (CI 54, 86)
			NPV	75% (CI 59, 91)
			NLR	0.67
Carter CJ et al, 1999, Canada	200 consecutive patients referred to diagnostic radiology department with ?DVT. Inpatients and outpatients	Prospective cohort	Prevalence	28%	No sample size calculation. Used single ultrasound as reference standard. Wide confidence intervals
			Sensitivity	87% (CI 80, 96)
			Specificity	79%
			NPV	94%
			NLR	0.16
Lennox AF et al, 1999, UK	200 consecutive patients referred to diagnostic radiology department with ?DVT. Inpatients and outpatients	Prospective cohort	Prevalence	23%	No sample size calculation. Incorrect test procedure likely to give falsely high sensitivities. No confidence intervals
			Sensitivity	91%
			Specificity	82%
			NPV	97%
			NLR	0.11
Farrell S et al, 2000, USA	173 consecutive patients referred to ED with ?DVT (48) or ?PE (125)	Prospective clinical trial	Prevalence	33%	Did not recruit all patients required. Used single ultrasound as reference standard. Wide confidence intervals
			Sensitivity	56% (CI 32, 81)
			NPV	77% (CI 62, 92)
			NLR	0.61 (CI 0.34, 1.11)
Van der Graaf F et al, 2000, Netherlands	112 outpatients referred to department	Prospective cohort	Prevalence	50%	Small patient numbers. No sample size calculation. Wide confidence intervals
			Sensitivity	80% (CI 66, 90)
			Specificity	94% (CI 83, 99)
			NPV	82% (CI 70, 91)
			Likelihood ratio for negative result (NLR)	0.21