Prognostic Value of D-Dimer in Stable Patients with Pulmonary Embolism

Abstract

Prognosis of pulmonary embolism (PE) based on levels ofD-dimer has shown mixed results, and data on in-hospital prognosis of stable patients are sparse. We assessed in-hospital prognosis in 292 stable patients with PE based on retrospective chart review using an arbitrarily selected value ofD-dimer ≥5000 ng/mL as cut-off level. In-hospital mortality from PE was 0% (0 of 222) withD-dimer <5000 ng/mL compared with 2.9% (2 of 70) withD-dimer ≥5000 ng/mL (P = .06). In-hospital all-cause mortality was 2.3% (5 of 222) withD-dimer <5000 ng/mL compared with 2.9% (2 of 70) withD-dimer ≥5000 ng/mL (NS). Markedly elevated levels ofD-dimer, therefore, did not indicate a high mortality from PE or all-cause mortality during hospitalization.

Levels ofD-dimer in patients with acute pulmonary embolism (PE) seem to relate to the severity of clot burden.^1,2 High levels may suggest a proximal location of the thrombus. ³ Prognosis of PE based on levels ofD-dimer has shown mixed results in mortality from PE^3–6 and we found only a single study in which in-hospital prognosis was assessed in stable patients. ⁵ In view of the mixed reports and sparse data on the value ofD-dimer in assessing prognosis of PE, we assessed our experience.

Methods

This was a retrospective study of stable patients hospitalized with acute PE who had measurements ofD-dimer. Patients were studied from March 2007 through May 2010 at St. Joseph Mercy Oakland Hospital, Pontiac Michigan, and from January 2004 through June 2008 at William Beaumont Hospital, Royal Oak, Michigan. The investigation was approved by the institutional review boards of both hospitals. Hospitalized patients with PE were identified by computer listings of International Classification of Diseases, Ninth Revision (ICD-9) discharge codes. Among 947 stable patients with PE,D-dimer was measured in 292 (31%).

The diagnosis of PE was made by computed tomographic (CT) pulmonary angiography in 280 patients (96%) and a high probability interpretation of the ventilation/perfusion lung scan in 12 patients (4%). Patients were excluded if they were hypotensive (systolic blood pressure <90 mm Hg) or on ventilatory support.

The mini Vitek Immuno Diagnostic Assay System (VIDAS) QuantitativeD-Dimer AssayD-Dimer Assay (Durham, North Carolina), an enzyme-linked fluorescent immunoassay, was performed at William Beaumont Hospital on the automated MINI VIDAS Instrument in 166 (57%) patients. Normal was <500 ng/mL. The Biosite TriageD-dimer flourescence immunoassay with the Triage MeterPLus (San Diego, California) was used for quantitative determination ofD-dimer in 126 (43%) patients at St Joseph Mercy Oakland Hospital. Normal was <400 ng/mL. D-dimer was obtained within 48 hours before or after the diagnosis of PE was made.

Cardiac troponin I (cTnI) was measured by Chemiluminescent assay, Access immunoassay system, Beckman Coulter, Fullerton, California, at St. Joseph Mercy Oakland Hospital, and ADVIA Centaur TnI-Ultra, Deerfield, Illinois, at the William Beaumont Hospital. Prior to October 2006 at William Beaumont Hospital, cTnI was defined as normal if ≤0.3 ng/mL. Afterward, at William Beaumont Hospital and at St. Joseph Mercy Oakland Hospital, cTnI was defined as normal if <0.05 ng/mL.

Creatine kinase isoenzyme MB (CK-MB) at William Beaumont Hospital was assayed in with 2-site sandwich immunoassay (ADVIA Centaur CK-MB, Deerfield, Illinois). Levels >5.0 ng/mL were defined as elevated (high). CK-MB at St. Joseph Mercy Oakland Hospital was assayed in 10 patients with a modified 2-site immunoenzymatic (sandwich) assay (Access immunoassay system, Beckman Coulter, Fullerton, California). Levels >6.3 ng/mL were defined as elevated (high).

Right ventricular (RV) dilatation by echocardiography was determined by a qualitative impression on the reports in 186 (85%) of 219 in whom echocardiograms were obtained. Right ventricular/left ventricular (RV/LV) ratios were measured in 33 (15%) of 219. An RV/LV ratio >1 was defined as RV dilatation. ⁷

Statistical analysis

We arbitrarily selected a D-dimer cut-off of 5000 ng/mL for comparisons of mortality with values above and below this level. Data were analyzed using SPSS Version 11.5 for Windows (SPSS Inc., Chicago, Illinois). Significant tests of equality of 2 proportions were carried out using the 2-tail Fisher exact test (http://www.graphpad.com/quickcalcs/contingency2.cfm); 95% confidence intervals (CI) were calculated using online software (http://www.hutchon.net/ConfidOR.htm). Student t-test was used for comparing means of continuous variables.

Results

Baseline characteristics of patients are shown in Table 1. None of the baseline characteristics differed significantly comparing those with D-dimer <5000 ng/mL with those with D-dimer ≥5000 ng/mL.

Table 1.

Baseline Characteristics According to D-dimer Level ^a

Variables	D-Dimer <5000 ng/mL	D-Dimer ≥5000 ng/mL
Age	63 ± 18	65 ± 14
Men	100/222 (45%)	31/70 (44%)
RV dilated	39/156 (25%)	24/63 (38%)
Elevated CK-MB	7/122 (6%)	4/37 (11%)
High Troponin	10/211 (5%)	4/59 (7%)

Abbreviation: RV, right ventricle.

^a All differences not significant.

In-hospital mortality from PE was 0% (0 of 222) with D-dimer <5000 ng/mL compared with 2.9% (2 of 70) with D-dimer ≥5000 ng/mL (P = .06; Table 1). In-hospital all-cause mortality was 2.3% (5 of 222) with D-dimer <5000 ng/mL compared with 2.9% (2 of 70) with D-dimer ≥5000 ng/mL (NS; Table 2).

Table 2.

Mortality and D-Dimer Level

	Mortality		Mortality		Probability
	D-dimer <5000 ng/mL		D-dimer ≥5000 ng/mL
	n/N (%)	95% CI	n/N (%)	95% CI
All cause	5/222 (2.3)	(1.0%-5.2%)	2/70 (2.9)	(0.8-9.8)	. 67
PE cause	0/222 (0)	(0%-1.7%)	2/70 (2.9)	(0.8-9.8)	. 06

Abbreviation: PE, pulmonary embolism.

Among patients with D-dimer ≥5000 ng/mL who had RV dilatation, mortality from PE was 4.2% (1 of 24) compared with 0% (0 of 39) among patients with D-dimer <5000 ng/mL who had RV dilatation (NS). Combinations of D-dimer level and RV dilatation are shown in Table 3.

Table 3.

Death From Pulmonary Embolism According to D-Dimer and Right Ventricular Size

Variable	PE-Mortality	PE-Mortality
	D-dimer ≥5000 ng/mL	D-dimer <5000 ng/mL
RV dilated	1/24 (4.2%)	0/39 (0%)
RV not dilated	1/39 (2.6%)	0/117 (0%)

Abbreviations: RV, right ventricle; PE, pulmonary embolism.

Discussion

The prevalence of in-hospital all-cause mortality did not differ significantly among patients with D-dimer ≥ 5000 ng/mL (2.9% mortality) and those with lower levels (2.3% mortality). In-hospital mortality from PE showed differences that almost reached statistical significance (0% vs 2.9%).

Previous investigators used cut-off levels of 6-fold higher than normal, ⁵ 3000 ng/mL, ³ 4200 ng/mL, ⁶ and 5500 ng/mL. We arbitrarily selected 5000 ng/mL.

The only previous investigation of in-hospital mortality with stable patients showed, as we showed, no difference in all-cause mortality or PE-related mortality ⁵ (Table 4). All-cause mortality, however, with longer follow-up of 15 days to 3 months, was shown to be higher in patients with elevated D-dimer levels.^3–6 Mortality from PE at 15 days to 3 months was found to be higher in one investigation ⁶ but not in another. Irrespective of whether patients with markedly elevated levels of D-dimer showed statistically significantly higher mortalities from PE, the mortality rates were insufficient to warrant thrombolytic or invasive therapy.

Table 4.

Mortality Related to D-Dimer

First Author (Year)	Number Patients	Included Patients	D-Dimer Levels Evaluated (ng/mL)	All-Cause Mortality	Pulmonary Embolism Mortality	Follow-up Period
Bova 2009 5	201	Stable PE	6-fold higher vs Normal	4/106 (3.8%) vs 0/95 (0%); NS	1/106 (0.9%) vs 0/95 (0%); NS	In-hospital
Lobo 2009 6	1707	All PE	>4200 vs <1050	30/426 (7.0%) vs 11/413 (2.7%);P = .004	17/426 (4.0%) vs 5/413 (1.2%);P = .016	15 days
Klok 2007 3	262	All PE *	>3000 vs <3000	Odds ratio 7.3; p = 0.017		3 months
Bova 2009 5	201	Stable PE	6-fold higher vs Normal	15/106 (14%) vs 3/95 (3%); P < 0.01		3-months
Aujesky 2006 4	366	Stable PE	>5500 vs <1500	8/88 (9.1%) vs 1/91 (1.1%); P < 0.04	3/88 (3.4%) vs 0/91 (0%); NS	3 months

Abbreviations: NS, not significant; PE, pulmonary embolism.

*Clinical probability “unlikely”.

Strengths of this investigation are that both all-cause mortality and PE-related mortality were assessed in stable patients, which, to our knowledge, had been done in only one previous investigation. A weakness is that relatively few patients were evaluated, although we evaluated a larger number than previously evaluated in stable patients.

In conclusion, markedly elevated levels of D-dimer did not indicate a high mortality from PE or all-cause mortality during hospitalization.