Serum CK-MB, COVID-19 severity and mortality: An updated systematic review and meta-analysis with meta-regression

Abstract

Objectives

We conducted a systematic review and meta-analysis with meta-regression of creatine kinase-MB (CK-MB), a biomarker of myocardial injury, in COVID-19 patients.

Methods

We searched PubMed, Web of Science and Scopus, for studies published between January 2020 and January 2021 that reported CK-MB, COVID-19 severity and mortality (PROSPERO registration number: CRD42021239657).

Results

Fifty-five studies in 11,791 COVID-19 patients were included in the meta-analysis. The pooled results showed that CK-MB concentrations were significantly higher in patients with high disease severity or non-survivor status than patients with low severity or survivor status (standardized mean difference, SMD, 0.81, 95% CI 0.61 to 1.01, p<0.001). The rate of patients with CK-MB values above the normal range was also significantly higher in the former than the latter (60/350 vs 98/1,780; RR ​= ​2.84, 95%CI 1.89 to 4.27, p<0.001; I² ​= ​19.9, p ​= ​0.254). Extreme between-study heterogeneity was observed (I² ​= ​93.4%, p<0.001). Sensitivity analysis, performed by sequentially removing each study and re-assessing the pooled estimates, showed that the magnitude and direction of the effect size was not modified (effect size range, 0.77 to 0.84). Begg's (p ​= ​0.50) and Egger's (p ​= ​0.86) t-tests did not show publication bias. In meta-regression analysis, the SMD was significantly and positively associated with the white blood count, aspartate aminotransferase, myoglobin, troponin, brain natriuretic peptide, lactate dehydrogenase, and D-dimer.

Conclusions

Higher CK-MB concentrations were significantly associated with severe disease and mortality in COVID-19 patients. This biomarker of myocardial injury might be useful for risk stratification in this group.

1. Introduction

Several clinical and demographic factors have been shown to be significantly associated with measures of coronavirus disease 2019 (COVID-19) severity, based on clinical and imaging findings and/or the need for aggressive single- and multi-organ support, and mortality [1,2]. The evidence of excessive inflammatory activity in severe COVID-19, captured during the early phases of the pandemic, prompted the investigation of the clinical role of specific biomarkers of inflammatory and immunomodulating pathways, particularly C-reactive protein (CRP), white blood cell count (WBC), neutrophils, lymphocytes, platelets, procalcitonin, ferritin, and serum amyloid A [[3], [4], [5], [6]]. Additional research has shown that patients with COVID-19 can also experience structural and functional abnormalities of specific organs and systems, e.g., cardiovascular, hematological, gastrointestinal, and neurological, in addition to the well-known respiratory compromise, characterized by the development of interstitial pneumonia and acute respiratory distress syndrome (ARDS) [[7], [8], [9]]. In particular, COVID-19 patients might present with subclinical or overt evidence of myocardial necrosis, which might manifest as acute coronary syndrome, myocarditis, arrhythmias, or heart failure [10]. While the exact mechanisms involved in the pathogenesis of cardiac complications in COVID-19 remain to be established, the presence of myocardial necrosis has been shown to be independently associated with more severe disease, transfer to the intensive care unit (ICU) and mortality [[11], [12], [13]]. Biomarkers of myocardial damage, particularly creatine kinase-MB (CK-MB) and troponin, have been increasingly investigated in COVID-19 patients in terms of their predictive capacity and potential to assist with clinical decisions [14]. A number of systematic reviews and meta-analyses on troponin and COVID-19 have been published [[15], [16], [17], [18]]. Similarly, meta-analyses have sought to critically appraise the available evidence regarding the clinical role of CK-MB [[19], [20], [21], [22]], with the largest meta-analysis identifying a total of 25 studies in 5,626 COVID-19 patients [21]. While CK-MB is less used in the diagnosis and the monitoring of myocardial necrosis in contemporary clinical practice since the advent of high-sensitivity troponin, additional retrospective and prospective studies have since been published on the associations between serum CK-MB concentrations, COVID-19 severity, and adverse outcomes. Furthermore, the assessment of CK-MB in COVID-19 patients might provide specific clinical information, independent of myocardial necrosis and cardiac complications, for early risk stratification in this group. We sought to address these issues by conducting an updated systematic review and meta-analysis of studies reporting serum CK-MB concentrations in COVID-19 patients with different disease severity, based on clinical guidelines or need for hospitalization, mechanical ventilation, or transfer to the ICU, and survival status during follow up. Additionally, a meta-regression analysis was performed to investigate possible associations between the effect size of the differences in CK-MB concentrations and a number of plausible clinical, demographic and biochemical factors.

2. Materials and methods

2.1. Search strategy, eligibility criteria, and study selection

We conducted a systematic literature search using the terms “CK-MB” or “creatine kinase MB” and “coronavirus disease 19” or “COVID-19”, in the electronic databases PubMed, Web of Science and Scopus, for peer-reviewed studies published from January 2020 to January 2021 that reported serum CK-MB concentrations in COVID-19 patients (PROSPERO registration number: CRD42021239657). The reference lists of the retrieved articles were also searched to identify additional studies. Eligibility criteria included: (a) studies reporting continuous data on CK-MB concentrations in COVID-19 patients, (b) articles investigating COVID-19 patients with different degrees of disease severity and/or survival status, (c) adult patients, (d) English language, (e) ≥10 COVID-19 patients, and (f) full-text available. Two investigators independently screened the abstracts. If relevant, the full articles were independently reviewed. The Newcastle-Ottawa scale was used to assess the quality of each study [23]. A score ≥6 indicated good quality when converting the scale to the Agency for Healthcare Research and Quality standards [24].

2.2. Statistical analysis

Standardized mean differences (SMD) and 95% confidence intervals (CIs) were calculated to build forest plots of continuous data and to evaluate differences in CK-MB concentrations between COVID-19 patients with low vs. high disease severity or survivor vs. non-survivor status. A p-value <0.05 was considered statistically significant. If studies reported CK-MB concentrations as median and interquartile range (IQR) or median and range, the corresponding means and standard deviations were estimated as previously described [25,26]. The Q-statistic was used to assess the heterogeneity of SMD values across studies with a significance level set at p<0.10. Inconsistency across studies was evaluated through the I² statistic, with I² values <25% indicating no heterogeneity, between 25 and 50% moderate heterogeneity, between 50 and 75% large heterogeneity, and >75% extreme heterogeneity [27,28]. A random-effect model was used to calculate the pooled SMD and corresponding 95% CIs in the presence of significant heterogeneity. In sensitivity analyses, the influence of each study on the overall effect size was assessed using the leave-one-out method [29]. The presence of publication bias was assessed using the Begg’s adjusted rank correlation test and the Egger’s regression asymmetry test at the p<0.05 level of significance [30,31], and further evaluated using the Duval and Tweedie “trim and fill” procedure. The latter, a funnel-plot-based method of testing and adjusting for publication bias, is a nonparametric (rank-based) data augmentation technique that increases the observed data, so that the funnel plot is more symmetric, and recalculates the pooled SMD based on the complete data [32]. To explore possible contributors to the between-study variance, we investigated the effects of several biologically and/or clinically plausible factors on the SMD by univariate meta-regression analysis. These factors included age, gender, clinical endpoint, month of recruitment commencement, diabetes, hypertension and cardiovascular disease, biomarkers of inflammation (CRP, WBC), liver damage (aspartate aminotransferase - AST, alanine aminotransferase - ALT, albumin), cardiac damage (myoglobin, troponin, brain natriuretic peptide - BNP), renal damage (serum creatinine), tissue damage and sepsis (lactate dehydrogenase - LDH, procalcitonin), and pro-thrombotic tendency (D-dimer). Statistical analyses were performed using Stata 14 (STATA Corp., College Station, TX, USA). The study was fully compliant with the PRISMA statement regarding the reporting of systematic reviews and meta-analyses [33].

3. Results

3.1. Literature search and study selection

Fig. 1 describes the flow chart of the screening process. We initially identified 661 studies. A total of 600 studies were excluded after the first screening because they were either duplicates or irrelevant. After a full-text revision of the remaining 61 articles, 6 were further excluded because they did not meet the inclusion criteria. Thus, 55 studies in 11,791 COVID-19 patients, 9,596 (48% males, mean age 54 years) with low severity or survivor status and 2,195 (62% males, mean age 66 years) with high severity or non-survivor status during follow up, were included in the meta-analysis (Table 1 ) [14,[34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86], [87].

Fig. 1.

Flow chart of study selection.

Table 1.

Characteristics of the selected studies.

				Mild disease or survivor				Severe disease or non-survivor
First Author,	Study		NOS	n	Age	Gender	CK-MB	n	Age	Gender	CK-MB
Country [reference]	design	Endpoint	(stars)		(Years)	(M/F)	U/L		(Years)	(M/F)	U/L
							(Mean ±SD)				(Mean ±SD)
Aladağ N et al.,	R	Survival status	7	35	68	22/13	29 ​± ​23	15	68	6/9	33 ​± ​25
Turkey [34]
Cao L et al.,	R	Survival status	7	78	68	33/45	0.8 ​± ​0.4a	22	81	12/10	4.9 ​± ​5.4a
China [35]
Chen FF et al.,	R	Survival status	7	577	63	297/280	1.1 ​± ​0.6a	104	73	65/39	3.7 ​± ​3.1a
China [36]
Chen J et al.,	R	Mechanical ventilation	7	68	67	38/30	36 ​± ​77	30	68	18/12	10 ​± ​13
China [37]
Deng M et al.,	R	Disease severity	7	53	35	24/29	10 ​± ​5	12	33	12/0	10 ​± ​13
China [38]
Deng P et al.,	R	Survival status	7	212	63	97/115	1.1 ​± ​0.6a	52	75	33/19	2.0 ​± ​0.5a
China [39]
Dong X et al.,	R	Survival status	7	65	54	30/35	24 ​± ​15	54	70	38/16	39 ​± ​29
China [40]
Du RH et al.,	R	ICU admission	5	58	68	38/20	3.7 ​± ​4.1a	51	79	36/15	5.4 ​± ​6.7a
China [41]
Feng X et al.,	P	Composite endpointb	7	94	63	58/36	18 ​± ​6	20	69	13/7	21 ​± ​14
China [42]
Feng Y et al.,	R	Disease severity	5	352	51	190/162	13 ​± ​4	124	60	81/43	16 ​± ​7
China [43]
Gao S et al.,	R	Survival status	5	175	70	79/96	9 ​± ​4	35	74	22/13	13 ​± ​8
China [44]
Gong J et al.,	R	Disease severity	5	161	45	72/89	12 ​± ​5	28	64	16/12	16 ​± ​17
China [45]
Guo H et al.,	R	Survival status	7	28	59	NR	6 ​± ​21a	46	72	NR	11 ​± ​49a
China [46]
Guo J et al.,	R	Survival status	6	43	60	22/21	11 ​± ​4	31	68	21/10	18 ​± ​7
China [47]
Han H et al.,	R	Disease severity	5	198	60	71/127	1.0 ​± ​0.6a	75	59	26/49	1.3 ​± ​1.1a
China [48]
He B et al.,	NR	Disease severity	5	32	42	15/17	0.6 ​± ​0.2a	21	57	13/8	1.6 ​± ​1.2a
China [49]
Hu J et al.,	R	Disease severity	7	130	63	58/72	13 ​± ​17	52	64	42/10	42 ​± ​93
China [50]
Hu X et al.,	R	Disease severity	7	175	41	80/95	9.2 ​± ​4.5	38	53	22/16	8.9 ​± ​4.4
China [51]
Jang JG et al.,	R	Disease severity	7	87	68	34/53	1.9 ​± ​1.8a	23	54	14/9	5.3 ​± ​4.1a
Korea [52]
Ji L et al.,	P	Composite endpointc	7	243	52	121/122	0.4 ​± ​0.2	37	71	20/17	14 ​± ​20
China [53]
Li N et al.,	R	Disease severity	7	103	61	48/55	0.8 ​± ​0.7a	35	67	23/12	2.7 ​± ​2.9a
China [54]
Li Y et al.,	R	Survival status	7	64	54	30/34	11.7 ​± ​3.7	37	72	23/14	22.3 ​± ​16.3
China [55]
Liu SL et al.,	R	Disease severity	5	194	43	91/103	9.3 ​± ​3.4	31	64	17/14	13.7 ​± ​7.4
China [56]
McRae MP et al.,	R	Survival status	5	117	63	52/65	5.5 ​± ​7.5a	43	73	30/13	8.6 ​± ​11.8a
USA [57]
Mertoglu C et al.,	R	ICU admission	7	532	48	306/226	16.4 ​± ​5.5	23	59	13/10	49 ​± ​52
Turkey [58]
Qin W et al.,	R	Survival status	7	239	63	113/126	0.9 ​± ​0.4a	23	69	10/13	2.0 ​± ​1.3a
China [59]
Tao Z et al.,	R	Disease severity	7	202	54	72/130	4.1 ​± ​5.5a	20	65	8/12	2.5 ​± ​2.2a
China [60]
Tuo H et al.,	R	Survival status	5	96	51	38/58	0.8 ​± ​0.6a	52	69	29/23	2.8 ​± ​2.3a
China [61]
Wang B et al.,	R	Survival status	7	54	62	40/14	24 ​± ​23	50	72	40/10	29 ​± ​22
China [62]
Wang C et al.,	R	Disease severity	6	35	38	17/18	16 ​± ​13	10	43	6/4	20 ​± ​7
China [63]
Wang D et al.,	R	Disease severity	5	72	44	29/43	13 ​± ​5	71	65	44/27	14 ​± ​8
China [64]
Wang F et al.,	R	Disease progression	7	253	41	109/144	0.6 ​± ​0.6a	70	60	45/25	0.7 ​± ​0.7a
China [65]
Wang G et al.,	NR	Disease severity	7	193	42	95/98	60 ​± ​55	16	54	10/6	8.4 ​± ​4.6
China [66]
Wang JH et al.,	R	Survival status	7	1074	61	502/572	0.8 ​± ​0.4a	61	74	43/18	2.5 ​± ​2.7a
China [67]
Wang K et al.,	P	Survival status	7	277	46	129/148	14 ​± ​4	19	66	11/8	21 ​± ​8
China [68]
Wang Z et al.,	R	Survival status	7	100	58	44/56	1.0 ​± ​0.5	56	72	32/24	6.4 ​± ​8.6
China [69]
Wu C et al.,	R	Presence of ARDS	6	117	48	68/49	15.3 ​± ​5.2	84	59	60/24	16.8 ​± ​5.6
China [70]
Xie J et al.,	NR	Disease progression	6	75	51	45/30	19.3 ​± ​5.9	29	66	18/11	22 ​± ​7.4
China [71]
Xie Y et al.,	R	Disease severity	5	38	61	14/24	0.6 ​± ​0.4	24	72	13/11	0.8 ​± ​0.5
China [72]
Xu B et al.,	R	Survival status	5	117	56	59/58	14 ​± ​6	28	73	17/11	21 ​± ​10
China [73]
Yang A et al.,	R	Disease severity	5	99	49	49/50	11 ​± ​10	15	60	7/8	12.1 ​± ​7.2
China [74]
Yang C et al.,	R	Survival status	7	145	57	77/68	13 ​± ​6	58	67	38/20	19 ​± ​7
Chin [75]
Yang J et al.,	R	Survival status	5	332	55	170/162	1.2 ​± ​1.0a	25	75	15/10	7 ​± ​6.7a
China [14]
Yao Q et al.,	R	Survival status	6	96	51	36/60	16 ​± ​5	12	65	7/5	23 ​± ​10
China [76]
Yu Z et al.,	R	Survival status	7	123	80	46/77	7.3 ​± ​3.7	18	84	11/7	8.8 ​± ​6.1
China [77]
Zeng Z et al.,	R	ICU admission	7	406	43	206/200	10.6 ​± ​5.5	55	60	33/22	12 ​± ​6.4
China [78]
Zhang C et al.,	R	Disease severity	7	56	44	24/32	0.5 ​± ​0.7	24	65	9/15	1.1 ​± ​1.2
China [79]
Zhang G et al.,	R	Disease severity	5	166	51	73/93	12.3 ​± ​3.7	55	62	35/20	22.3 ​± ​15.6
China [80]
Zhang JJ et al.,	R	Survival status	7	240	53	119/121	1.2 ​± ​1.0	49	69	25/14	2.3 ​± ​1.9
China [81]
Zhang Q et al.,	R	Disease severity	7	47	61	18/29	9.3 ​± ​3.0	27	72	18/9	16.3 ​± ​8.9
China [82]
Zhang XB et al.,	R	Survival status	7	410	53	219/191	7.1 ​± ​7.6	22	66	11/11	19.4 ​± ​19.6
China [83]
Zhao Y et al.,	R	Disease severity	7	336	43	145/191	0.8 ​± ​0.4	81	56	53/28	1.1 ​± ​0.4
China [84]
Zhou C et al.,	R	Disease severity	7	95	35	38/57	6.9 ​± ​3.7	28	40	17/11	8.0 ​± ​2.0
China [85]
Zhou J et al.,	NR	ICU admission	5	156	40	75/81	11.1 ​± ​6.2	45	57	27/18	9.3 ​± ​4.7
China [86]
Zhu Y et al.,	R	Survival status	5	73	73	39/34	1.8 ​± ​1.3	29	62	19/20	7.2 ​± ​8.7
China [87]

Abbreviations: ICU, intensive care unit; NOS, Newcastle-Ottawa quality assessment scale for case–control studies; NR, Not Reported; P, prospective; R, retrospective; ARDS, acute respiratory distress syndrome.

^a-ng/mL or μg/L.

^b-survival status on discharge, disease severity, and mechanical ventilation.

^c-survival status on discharge, presence of ARDS.

Two studies were conducted in Turkey [34,58], 1 in the USA [57], 1 in Korea [52], and the remaining 51 in China [14,[35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51],[53], [54], [55], [56],[59], [60], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86], [87]]. Three studies were prospective [42,53,68], 47 retrospective [[34], [35], [36], [37], [38], [39], [40], [41],[43], [44], [45], [46], [47], [48],[50], [51], [52],[54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65],67,[69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85],87], whereas the remaining 5 did not report information regarding the study design [14,49,66,71,86]. Clinical endpoints included disease severity based on current clinical guidelines in 21 studies [38,43,45,[48], [49], [50], [51], [52],54,56,60,63,64,66,72,74,79,80,82,84,85], admission to ICU in 4 [41,58,78,86], clinical progress in 2 [65,71], presence of ARDS in 1 [70], need for mechanical ventilation in 1 [37], and combined clinical outcomes in 2 [42,53], and survival status in 24 [14,[34], [35], [36],39,40,44,46,47,55,57,59,61,62,[67], [68], [69],73,[75], [76], [77],81,83,87]. Only 1 study reported the highest CK-MB serum concentrations during hospitalization [61], whereas the remaining 54 reported a single CK-MB concentration within the first 24–48 ​h from admission.

3.2. Meta-analysis

The overall SMD in CK-MB concentrations between COVID-19 patients with low vs. high disease severity or survivor vs. non-survivor status during follow up is shown in Fig. 2 . In 5 studies, patients with high severity or non-survivor status had lower CK-MB concentrations when compared to those with low severity or survivor status (mean difference range, -0.07 to -0.97) [37,51,60,66,86], although only 1 study reported a statistically significant difference [66]. In 1 study, no difference was observed (mean difference 0.00) [38], whereas in the remaining 49 studies the CK-MB concentrations were lower in patients with low severity or survivor status (mean difference range, 0.11 to 2.91), with a non-significant difference in 17 [34,38,41,42,46,57,[62], [63], [64], [65],[70], [71], [72],74,77,78,85]. The pooled results confirmed that the CK-MB concentrations were statistically significantly higher in patients with high disease severity or non-survivor status during follow up (SMD 0.81, 95% CI 0.61 to 1.01, p<0.001) (Fig. 2). An extreme heterogeneity between studies was observed (I² ​= ​93.4%, p<0.001). A sub-group of 11 studies reported the proportion of patients with CK-MB concentrations above the normal range in COVID-19 patients with high severity or non-survivor status vs. those with low severity or survivor status during follow-up [38,41,44,48,51,52,58,76,77,81,86]. The rate of participants with CK-MB values above the normal range was statistically significantly higher in the former than the latter (60/350 vs 98/1,780; RR ​= ​2.84, 95%CI 1.89 to 4.27, p<0.001; I² ​= ​19.9, p ​= ​0.254) (Fig. 3 ).

Fig. 2.

Forest plot of studies reporting CK-MB concentrations in patients with COVID-19.

Fig. 3.

Forest plot of studies reporting the proportion of patients with CK-MB concentrations above the normal range in COVID-19 patients with high severity or non-survivor status vs. those with low severity or survivor status during follow-up.

Sensitivity analysis, performed by sequentially removing each study and re-assessing the pooled estimates, showed that the magnitude and the direction of the effect size were not influenced (effect size range, between 0.77 and 0.84) (Fig. 4 ). CK-MB concentrations remained statistically significantly higher (SMD 0.72, 95% CI 0.54 to 0.90, p<0.001; I² ​= ​90.2%, p<0.001) in patients with high severity or non-survivor status after excluding 5 relatively large studies that accounted for nearly 28% of the overall sample size [36,58,67,78,83].

Fig. 4.

Sensitivity analysis of the association between CK-MB and COVID-19 disease. The influence of individual studies on the overall standardized mean difference (SMD) is shown. The middle vertical axis indicates the overall SMD and the two vertical axes indicate the 95% confidence intervals (CIs). The hollow circles represent the pooled SMD when the remaining study is omitted from the meta-analysis. The two ends of each broken line represent the 95% CIs.

No publication bias was observed with the Begg's (p ​= ​0.50) and Egger's (p ​= ​0.86) t-tests. Accordingly, the trim-and-fill analysis showed that no study was missing or should be added (Fig. 5 ).

Fig. 5.

Funnel plot of studies investigating disease severity of survival status after trimming and filling. Dummy studies and genuine studies are represented by enclosed circles and free circles, respectively.

3.3. Meta-regression analysis

In univariate meta-regression analysis, the SMD was statistically significantly and positively associated with the WBC (t ​= ​3.51, p ​= ​0.001), AST (t ​= ​2.81, p ​= ​0.007), myoglobin (t ​= ​2.91, p ​= ​0.01), troponin (t ​= ​3.83, p ​= ​0.001), BNP (t ​= ​2.67, p ​= ​0.02), LDH (t ​= ​2.40, p ​= ​0.02), and D-dimer (t ​= ​2.47, p ​= ​0.02) (Table 2 ). By contrast, no statistically significant associations were observed between the SMD and age (t ​= ​0.43, p ​= ​0.67), gender (t ​= ​−0.54, p ​= ​0.59), month of recruitment commencement (t ​= ​1.44, p ​= ​0.15; Fig. 6 ), CRP (t ​= ​1.17, p ​= ​0.25), ALT (t ​= ​−0.73, p ​= ​0.47), creatinine (t ​= ​1.17, p ​= ​0.25), procalcitonin (t ​= ​1.75, p ​= ​0.10), albumin (t ​= ​−1.53, p ​= ​0.13), diabetes (t ​= ​0.56, p ​= ​0.58), hypertension (t ​= ​−0.80, p ​= ​0.43) and cardiovascular disease (t ​= ​0.59, p ​= ​0.56) (Table 2).

Table 2.

Univariate meta-regression analysis between effect size and possible contributors to heterogeneity.

	Number of studies	t	p-value	Explained heterogeneity
Age	54	0.43	0.67	0%
Gender	54	−0.54	0.59	0%
Time of recruitment commencement	54	1.44	0.15	3%
White blood cell count	44	3.51	0.001	22%
C-reactive protein	46	1.17	0.25	1%
Procalcitonin	23	1.75	0.10	9%
Aspartate aminotransferase	47	2.81	0.007	15%
Alanine aminotransferase	48	−0.73	0.47	0%
Albumin	37	−1.53	0.13	4%
D-dimer	40	2.47	0.02	13%
Troponin	26	3.83	0.001	39%
Myoglobin	17	2.91	0.01	36%
Brain natriuretic peptide	19	2.67	0.02	29%
Creatinine	43	1.17	0.25	5%
Lactate dehydrogenase	34	2.40	0.02	15%
Diabetes	37	0.56	0.58	0%
Cardiovascular disease	34	−0.80	0.43	0%
Hypertension	39	0.59	0.56	0%

Fig. 6.

Bubble plot reporting univariate meta-regression analysis between the month of recruitment commencement and SMD. Each study is represented by a circle (bubble) with the size area proportional to the study precision (the inverse of its within-study variance).

In sub-group analysis, the pooled SMD in studies assessing disease severity (SMD 0.52, 95% CI 0.32 to 0.54, p<0.001; I² ​= ​84.6, p<0.001) was statistically significantly lower than that observed in studies assessing survival status (SMD 1.18, 95% CI 0.89 to 1.46, p<0.001; I² ​= ​92.0, p<0.001; t ​= ​3.38, p ​= ​0.002) (Fig. 7 ). However, the between-study variance remained extreme regardless of the type of endpoint studied.

Fig. 7.

Forest plot of studies examining CK-MB concentrations in patients with COVID-19 according to disease severity or survival status. The diamond represents the point estimate and confidence intervals after combining and averaging all individual studies. The vertical line through the vertical points of the diamond represents the point estimate of the averaged studies.

We further sought to identify additional, more homogeneous, study sub-groups according to endpoint reported, study design (retrospective or prospective), country, and assay type (measurement of CK-MB on the basis of activity or mass). In particular, a sub-group of 13 retrospective studies conducted in China [40,44,47,55,62,69,73,[75], [76], [77],81,83,87] assessed survival and measured CK-MB using an assay based on activity evaluation. Funnel plot analysis showed that 3 of these studies were located outside of the plot (Fig. 8 A) [62,77,83]. After removing these studies, the effect size of the remaining 10 studies confirmed that the CK-MB concentrations were statistically significantly higher in patients with high disease severity or non-survivor status during follow up (SMD 0.97, 95% CI 0.84 to 1.09, p<0.001) (Fig. 8B). However, a substantially lower heterogeneity between studies was observed (I² ​= ​0.0%, p ​= ​0.72).

Fig. 8.

Funnel plot of a sub-group of 13 studies that were homogeneous for endpoint, country, study design and assay (A). Forest plot of CK-MB concentrations after removing the three studies that were outside the funnel plot (B).

4. Discussion

In this updated systematic review and meta-analysis, the serum concentrations of CK-MB, measured within 24–48 ​h from admission in virtually all studies, were statistically significantly higher in patients with COVID-19 who had a more severe clinical picture, based on available guidelines, disease progress, need for mechanical ventilation, or transfer to ICU, or did not survive, during follow up. The magnitude of the observed SMD value suggests the presence of a clinically relevant between-group difference in CK-MB concentrations [88]. Whilst the study heterogeneity was extreme, the sequential omission of individual studies did not exert tangible effects on the SMD value. Furthermore, there was no evidence of publication bias according to the Begg’s and Egger’s t-tests and the trim and fill analysis. Previous systematic reviews and meta-analyses have reported associations between serum CK-MB concentrations and COVID-19 severity and mortality [[19], [20], [21], [22]]. Our meta-analysis has captured more than double the number of studies, 55 vs. 25, and participants, 11,791 vs. 5,626, than the largest previously published meta-analysis [21]. Furthermore, we investigated possible associations between the observed SMD and a number of biologically and clinically plausible factors. Using meta-regression analysis, there were statistically significant and positive associations between the SMD and the WBC and the concentrations of AST, myoglobin, troponin, BNP, LDH, and D-dimer.

CK-MB, one of the three CK isoenzymes, is present in high concentrations in the myocardium although it can also be detected in the brain and the skeletal muscle [89]. The release of CK-MB in the circulation, and its consequent increase in serum concentrations, has been used for many years to diagnose myocardial necrosis and its clinical manifestations, e.g., myocardial infarction, until more sensitive and specific biomarkers, i.e., high-sensitivity troponin, were introduced in clinical practice [89,90]. The results of our meta-analysis might primarily reflect the presence of significant myocardial damage in COVID-19 patients with more compromised clinical presentations and high-risk of mortality. However, the associations observed in meta-regression indicate that the clinical information provided by CK-MB might complement, rather than duplicate, that of other cardiac biomarkers. While, not surprisingly, the CK-MB SMD was statistically significantly associated with troponin, myoglobin, and BNP, the additional associations observed suggest that the elevations of CK-MB in high-risk COVID-19 patients can also reflect a state of excess inflammation (WBC), single- and multi-organ damage (AST and LDH), and pro-thrombotic tendency (D-dimer). Notably, these biomarkers have, in turn, shown significant associations with COVID-19 severity and adverse outcomes [16,91].

4.1. Limitations of the study

The extreme heterogeneity observed between the studies, together with the exclusion of articles written in non-English language, e.g., Chinese, is a limitation of our meta-analysis. However, the trend and magnitude of the reported differences in CK-MB were maintained, in presence of a significantly reduced heterogeneity, in a sub-group of 10 studies that were homogeneous for endpoint, study design, country, and assay used. In addition, it is important to emphasize that there was no evidence of publication bias and that the overall effect size was not affected in sensitivity analyses. A number of unreported factors might have contributed to the heterogeneity, particularly the coexistence of rhabdomyolysis, reported in COVID-19 patients and a common cause of increased serum CK-MB concentrations [[92], [93], [94]], and the fact that virtually all selected studies reported a single measurement of CK-MB rather than serial assessments.

5. Conclusions

In conclusion, our updated systematic review and meta-analysis with meta-regression has shown that higher serum CK-MB concentrations are significantly associated with worse clinical status and reduced survival in patients with COVID-19. Further research is warranted to establish the clinical utility of this biomarker, in conjunction with other patient characteristics, for early risk stratification and acute management in this group.

Financial disclosure

The authors have no funding to disclose.

The Author contribution

Study design: Angelo Zinellu.

Data collection: Angelo Zinellu.

Statistical analysis: Angelo Zinellu, Salvatore Sotgia.

Data interpretation: Angelo Zinellu, Salvatore Sotgia, Alessandro Fois, Arduino Mangoni.

Manuscript preparation: Angelo Zinellu, Salvatore Sotgia, Alessandro Fois, Arduino Mangoni.

Literature search: Angelo Zinellu, Arduino Mangoni.

Funds collection: n/a.

Declaration of competing interest

The authors declare no conflict of interests.