An updated meta‐analysis on the relationship between D‐dimer levels and severity of coronavirus disease 2019

1.

Dear Editors,

The outbreak of coronavirus disease 2019 (COVID‐19) is a global health emergency, which requires us to face and resolve this difficult problem together. As the epidemic continues to deteriorate, clinical and laboratory features are receiving considerable attention. ¹ It has been shown that many laboratory abnormalities might predict adverse outcomes including increased D‐dimer values. ² However, there is still plenty of evidence to prove this. Dr Lippi et al ³ conducted a meta‐analysis based on 4 studies to explore the relationship between D‐dimer levels and the severity of COVID‐19 by calculating weighted mean difference (WMD) and 95% confidence interval (CI). The results by Lippi et al suggested that D‐dimer values are higher in COVID‐19 patients with severe disease than in those without (WMD = 2.97 mg/L, 95% CI = 2.47‐3.46 mg/L). However, too few studies were included and the sample sizes were limited. Based on the above points, we have re‐searched, included new studies, and conducted an updated meta‐analysis.

This meta‐analysis was performed strictly according to the Preferred Reporting Items for Systematic Review and Meta‐analysis (PRISMA) statement (Table S1). ⁴ An electronic search was applied to PubMed, Web of Science, and medRxiv (https://connect.medrxiv.org/relate/content/181) databases, using the following search terms: “coronavirus” OR “2019‐nCoV” OR “SARS‐CoV‐2” OR “COVID‐19” AND “laboratory,” between January 1, 2020, and April 7, 2020. We screened the titles and abstracts of the studies and then enrolled the studies that documented the characteristic of D‐dimer levels in non‐severe and severe patients with COVID‐19 by reading full texts. Severe COVID‐19 patients included those with acute respiratory distress syndrome, admitted to the intensive care unit, or died. ³ , ⁵ Because of the existence of unit differences and high heterogeneity, the standardized mean difference (SMD) with 95% CI and the odds ratio (OR) with the corresponding 95% CI were calculated in severe patients versus non‐severe patients through a random‐effects model. ⁶ We estimated the mean and standard deviation according to Wan et al ⁷ when sample size, median, and interquartile range (IQR) were provided. All calculations were undertaken in STATA 11.2. An 11‐item checklist recommended by the Agency for Healthcare Research and Quality (AHRQ) was applied to assess the quality of included studies. ⁸

Initially, 1930 studies were identified. As a result, 21 studies with 3657 patients were included in the synthesis. All the patients in the included studies were from China. The units of D‐dimer were varied in the eligible studies, including mg/L, mg/mL, μg/mL, ng/mL, and μg/L. The quality scores ranged from 4 to 8. All studies were considered to be of high or moderate quality (Table S2). The number of severe COVID‐19 cases ranged from 4 to 173. Among 21 eligible studies, 19 studies with 2455 patients reported data about D‐dimer levels, while only 5 studies with 1136 patients provided the number of patients who had elevated D‐dimer levels between severe and non‐severe COVID‐19 patients, of which 3 defined increased D‐dimer level as D‐dimer > 0.5 mg/L, one as D‐dimer > 0.243 μg/mL, and one as undefined. The main features of the included studies are shown in Table 1.

Table 1.

Characteristics of the included studies

Author	Location	Age	Male	Quality score a	Unit	Severe			Non‐Severe
						n	D‐dimer	Increased D‐dimer	n	D‐dimer	Increased D‐dimer
Wang D	China	56 (median)	75 (54.3)	8	mg/mL b	36	414 (191‐1324)	NR	102	166 (101‐285)	NR
Huang C	China	49 (median)	30 (73.2)	6	mg/L	13	2.4 (0.6‐14.4)	NR	28	0.5 (0.3‐0.8)	NR
Cai Q	China	47 (median)	149 (51.6)	6	mg/L	58	0.96 (0.54‐1.925)	39/58	240	0.35 (0.25‐0.52)	60/240
Liu Y	China	55 (median)	59 (54.1)	7	mg/mL b	53	940 (470‐1905)	NR	56	370 (250‐650)	NR
Li J	China	45.1 (mean)	9 (52.9)	6	mg/mL b	12	250.9 ± 154.4	NR	5	105.7 ± 57.3	NR
Liu J	China	48.7 (mean)	15 (37.5)	7	mg/L	13	0.9 (0.7‐1.5)	NR	27	0.4 (0.2‐0.8)	NR
Luo X	China	57 (median)	150 (50.3)	8	mg/L	84	4.59 (0.95‐17.14)	NR	214	0.50 (0.29‐1.10)	NR
Liu M	China	35 (mean)	10 (33.3)	4	mg/L	4	1.54 ± 1.22	NR	26	0.26 ± 0.08	NR
Wan S	China	47 (median)	72 (53.3)	5	mg/L	40	0.6 (0.4‐1.1)	NR	95	0.3 (0.2‐0.5)	NR
Han H	China	NR	48 (51.1)	4	mg/L	35	19.11 ± 35.48	NR	49	2.14 ± 2.88	NR
Mao L	China	52.7 (mean)	87 (40.7)	7	mg/L	88	0.9 (0.1‐20.0)	NR	126	0.4 (0.2‐8.7)	NR
Zhang J	China	57 (median)	71 (50.7)	7	μg/mL	58	0.4 (0.2‐2.4) c	23/38	82	0.2 (0.1‐0.3) c	12/43
Mo P	China	54 (median)	86 (55.5)	7	ng/mL	85	213 (126‐447)	NR	70	178 (100‐249)	NR
Gao Y	China	NR	26 (60.5)	6	μg/L	15	490 (290‐910) b	NR	28	210 (190‐270) b	NR
Qian G	China	50 (median)	37 (40.7)	8	ng/mL	9	450 (160‐485)	NR	82	300 (106‐400)	NR
Wu C	China	51 (median)	128 (63.7)	8	μg/mL	84	1.16 (0.46‐5.37)	NR	117	0.52 (0.33‐0.93)	NR
Huang H	China	44.87 (mean)	63 (50.4)	6	μg/L	32	1760 (1297.5‐3265)	NR	93	780 (560‐1050)	NR
Tang N	China	54.1 (mean)	98 (53.6)	7	μg/mL	21	2.12 (0.77‐5.27)	NR	162	0.61 (0.35‐1.29)	NR
Zhou F	China	56 (median)	119 (62.3)	8	μg/mL	54	5.2 (1.5‐21.1)	50/54	137	0.6 (0.3‐1.0) c	67/118
Guan W	China	47 (median)	637 (58.1)	8	mg/L	173	NR	65/109	926	NR	195/451
Li Y	China	51 (median)	12 (48.0)	8	NA	9	NR	5/9	16	NR	6/16

All values are n (%), median (IQR), mean ± SD, or n/total.

Abbreviation: NR, not reported.

^a Quality scores obtained by using an 11‐item checklist suggested by the Agency for Healthcare Research and Quality.

^b Corrections according to a paper accepted for publication (Favaloro EJ, Thachil J. Reporting of D‐dimer data in COVID‐19: some confusion and potential for misinformation. Clin Chem Lab Med, 2020).

^c Data missing for patients.

The pooled SMD indicated that the D‐dimer levels of severe patients were higher than that of non‐severe patients (SMD = 0.97, 95% CI = 0.77‐1.17; I ² = 75.7%, P < .001; random‐effects model) (Figure 1A). As Figure 1B showed that D‐dimer elevation was more likely to be detected in severe COVID‐19 patients (OR = 3.27, 95% CI = 1.64‐6.51; I ² = 68.9%, P = .022; random‐effects model). The similar results were observed in non‐survival COVID‐19 patients (OR = 9.51, 95% CI = 3.23‐28.06; random‐effects model).

Figure 1.

A, Forest plot of standardized mean difference (SMD) in D‐dimer values between COVID‐19 patients with or without severe disease and (B) forest plot of odds ratio (OR) for the relationship between D‐dimer levels and severity of COVID‐19: subgroup analysis by COVID‐19 outcomes

Although the definitions of severe COVID‐19 were different across these included studies (Table S3), which may result in higher heterogeneity, our study found that elevated D‐dimer was an important feature of severe COVID‐19 patients. Elevation of D‐dimer is a marker of thrombosis and is also one of the diagnostic criteria for disseminated intravascular coagulation (DIC). Thrombosis and/or DIC may occur in many severe COVID‐19–infected patients. ⁹ Thus, corresponding timely treatment for these patients on this basis is indicated, for example, anticoagulant therapy if appropriate.

CONFLICTS OF INTEREST

All authors report that they have no potential conflict of interest.

AUTHOR CONTRIBUTIONS

Li Shi collected, extracted, and analyzed the data of eligible studies and wrote the manuscript; Ying Wang took part in the data extraction and statistical analysis; Yadong Wang contributed to statistical analysis and revision of manuscript; Guangcai Duan helped to revise the manuscript; Haiyan Yang designed the study; and each author has read and approved the final manuscript.

Supporting information

Appendix