Search results
A total of 495 studies were identified as potentially relevant literature reports. There were no additional studies identified through other sources. We got 405 articles when the duplicate was removed. By scanning the title and abstract, 363 studies were excluded according to the eligibility criteria. Another 22 articles were further excluded by reading the full text. Ultimately, 20[9, 15–33] articles were eligible for data extraction and meta-analysis. The searching process is shown in Fig. 1.
Characteristics Of Included Studies
The characteristics of the 20 included studies are summarized in Table 1. 2099 individuals were incorporated into our trial of whom 591 UC patients with CMV infection or reactivation and 1508 UC patients are CMV negative.
Table 1
The basic characteristics and quality evaluation of the included studies.
Author | Year | Country | Study design | Study Group | Gender | Age |
CMV+ | CMV- | CMV+(M/F) | CMV-(M/F) | CMV+ | CMV- |
Matsuoka | 2007 | Japan | Retrospective | 25 | 23 | 12/13 | 17/6 | 40.0(15–73) | 33.0(19–64) |
Omiya | 2010 | Japan | Prospective | 7 | 13 | NA | NA | NA | NA |
Suzuki | 2010 | Japan | Retrospective | 15 | 58 | 8/7 | 26/32 | 42 (17–65) | 34 (14–81) |
Criscuoli | 2011 | Italy | Prospective | 28 | 57 | 19/9 | 31/26 | NA | NA |
Kim | 2012 | Korea | Prospective | 31 | 41 | 20/11 | 28/13 | 43.1 ± 15.4 | 43.4 ± 14.1 |
Inokuchi | 2014 | Japan | Retrospective | 40 | 78 | 26/14 | 31/47 | 45 (14–79) | 36 (12–87) |
Chun | 2015 | Korea | Retrospective | 12 | 31 | 6/6 | 14/17 | NA | NA |
McCurdy | 2015 | USA | Retrospective | 45 | 139 | NA | NA | NA | NA |
Pillet | 2015 | France | Prospective | 39 | 70 | 28/11 | 44/26 | 51.2 ± 17.0 | 46.5 ± 18.0 |
Hirayama | 2016 | Japan | Retrospective | 34 | 115 | 19/15 | 64/51 | 42.3 ± 14.4 | 29.0 ± 14.4 |
Lee | 2016 | Korea | Retrospective | 50 | 99 | 30/20 | 50/49 | 45 ± 14.75 | 42 ± 10.33 |
ORMECI | 2016 | Turkey | Retrospective | 8 | 35 | 6/2 | 17/18 | NA | NA |
Zagórowicz | 2016 | Poland | Retrospective | 33 | 62 | 22/11 | 39/23 | NA | NA |
Henmi | 2018 | Japan | Retrospective | 26 | 60 | 15/11 | 38/22 | 55.0 ± 13.2 | 41.8 ± 16.2 |
Levin | 2017 | USA | Retrospective | 13 | 15 | 6/7 | 12/3 | 38 ± 19 | 41 ± 19 |
Nowacki | 2018 | Germany | Retrospective | 34 | 205 | 25/9 | 108/97 | 37 ± 17.8 | 36 ± 18.9 |
Feng | 2016 | China | Retrospective | 31 | 60 | 18/13 | 39/21 | NA | NA |
Xu | 2016 | China | Retrospective | 49 | 143 | 28/21 | 82/61 | 39.15 ± 2.84 | 38.32 ± 2.94 |
Li | 2019 | China | Retrospective | 34 | 91 | 21/13 | 49/42 | 43.08 ± 3.07 | 42.41 ± 2.36 |
Zhang | 2018 | China | Retrospective | 37 | 113 | 23/14 | 61/52 | 43.17 ± 3.08 | 42.45 ± 2.55 |
Study Quality And Risk Of Bias
The quality of the included studies was evaluated by the scale of NOS. NOS scores for all studies were greater than or equal to 7(Table 2).
Table 2
The Newcastle-Ottawa Scale (NOS) assessment of the included studies
Study | Year | Selection | Comparability | Exposure | Total |
Matsuoka | 2007 | 4 | 3 | 2 | 9 |
Omiya | 2010 | 4 | 2 | 2 | 8 |
Suzuki | 2010 | 4 | 2 | 3 | 9 |
Criscuoli | 2011 | 4 | 2 | 2 | 8 |
Kim | 2012 | 4 | 2 | 3 | 9 |
Inokuchi | 2014 | 3 | 3 | 2 | 8 |
Chun | 2015 | 4 | 2 | 3 | 9 |
McCurdy | 2015 | 4 | 2 | 3 | 9 |
Pillet | 2015 | 4 | 2 | 3 | 9 |
Hirayama | 2016 | 3 | 2 | 3 | 8 |
Lee | 2016 | 4 | 2 | 3 | 9 |
ORMECI | 2016 | 4 | 2 | 2 | 8 |
Zagórowicz | 2016 | 3 | 2 | 2 | 7 |
Henmi | 2018 | 3 | 2 | 3 | 8 |
Levin | 2017 | 4 | 2 | 3 | 9 |
Nowacki | 2018 | 4 | 1 | 3 | 8 |
Feng | 2016 | 4 | 1 | 3 | 8 |
Xu | 2016 | 4 | 2 | 2 | 8 |
Li | 2019 | 3 | 2 | 2 | 7 |
Zhang | 2018 | 4 | 1 | 3 | 8 |
Publication Bias
The details of Egger tests were presented in Table 3. The Egger tests did not indicate publication bias in all the outcomes.
Table 3
Results of Egger test for outcomes
| Number of Studies | Std_Eff | t | P>| t | | 95%CI |
Severe UC | 10 | Slope | -0.24 | 0.813 | -1.521 | 1.231 |
| | Bias | 0.88 | 0.404 | -1.866 | 4.172 |
Pancolitis | 14 | Slope | 1.66 | 0.123 | -0.206 | -1.442 |
| | Bias | 0.17 | 0.868 | -1.442 | 1.686 |
Age of onset | 8 | Slope | 0.87 | 0.419 | -22.170 | 9.180 |
| | Bias | -0.42 | 0.690 | -12.965 | 9.180 |
Disease duration | 9 | Slope | -2.27 | 0.057 | -10.227 | 0.201 |
| | Bias | 1.71 | 0.130 | -1.252 | 7.857 |
Glucocorticoid | 14 | Slope | 1.28 | 0.224 | -0.477 | 1.841 |
| | Bias | 1.40 | 0.186 | -0.688 | 3.184 |
Immunosuppressant | 15 | Slope | 0.48 | 0.642 | -1.186 | 1.854 |
| | Bias | 0.50 | 0.623 | -2.391 | 3.598 |
Azathioprine | 5 | Slope | 0.41 | 0.706 | -2.628 | 3.415 |
| | Bias | -0.09 | 0.937 | -5.430 | 5.144 |
Infliximab | 7 | Slope | 0.87 | 0.423 | -1.860 | 3.770 |
| | Bias | -0.24 | 0.821 | -5.632 | 4.675 |
5-ASA | 9 | Slope | -1.66 | 0.140 | -2.285 | 0.399 |
| | Bias | 0.99 | 0.354 | -1.554 | 3.803 |
Methodological Quality Assessment
The total qualities of the evidence were low for the severe UC, glucocorticoid, immunosuppressant as well as very low for pancolitis and age of onset (Table 4).
Table 4
Risk factors of cytomegalovirus infection or reactivation in UC patients
Patient or population: UC patients with definite diagnosis Settings: Risk factors of cytomegalovirus infection or reactivation Intervention: UC patients with OR without CMV infection or reactivation |
Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | No of Participants (studies) | Quality of the evidence (GRADE) | Comments |
Assumed risk | Corresponding risk |
| Control | New Comparison | | | | |
Severe UC | Study population | OR 1.46 (1.1 to 1.93) | 1169 (10 studies) | ⊕⊕⊝⊝ low | |
533 per 1000 | 625 per 1000 (557 to 688) |
Moderate |
531 per 1000 | 623 per 1000 (555 to 686) |
Pancolitis | Study population | OR 1.77 (1.35 to 2.32) | 1516 (14 studies) | ⊕⊝⊝⊝ very low1 | |
625 per 1000 | 747 per 1000 (693 to 795) |
Moderate |
657 per 1000 | 772 per 1000 (721 to 816) |
Glucocorticoid | Study population | OR 3.75 (2.79 to 5.03) | 1488 (14 studies) | ⊕⊕⊝⊝ low | |
497 per 1000 | 787 per 1000 (734 to 832) |
Moderate |
514 per 1000 | 799 per 1000 (747 to 842) |
Immunosuppressant | Study population | OR 1.92 (1.49 to 2.47) | 1662 (15 studies) | ⊕⊕⊝⊝ low | |
229 per 1000 | 363 per 1000 (307 to 423) |
Moderate |
175 per 1000 | 289 per 1000 (240 to 344) |
Age of onset | | The mean age of oneset in the intervention groups was 6.21 higher (2.55 to 9.87 higher) | | 965 (8 studies) | ⊕⊝⊝⊝ very low | |
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; |
GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. |
1 The statistical test for heterogeneity shows a low P value. |
Outcomes Of The Meta-analysis
Severe UC
All the UC patients were evaluated by modified Mayo scoring system and modified Truelove and Witts scoring system. We extracted the data of amount of severe UC and mild to moderate UC patients from 10 included articles. The results show that the risk of CMV infection or reactivation in severe UC patients was 1.465 times higher than that in mild to moderate UC patients. (heterogeneity P = 0.858, I2 = 0.0%, OR = 1.465, 95% CI: 1.107 to 1.939, P = 0.008; Fig. 2).
Pancolitis
14 articles reported on colonoscopy in patients with UC to determine the extent of intestinal lesions. There is no significant heterogeneity in the statistical results of the pooled literature (I2 = 0.0%, P = 0.507). The result of the fixed effect model showed that the pancolitis could increase the risk of CMV infection or reactivation (OR = 2.108, 95% CI: 1.586 to 2.800, P = 0.0001; Fig. 3).
Age of onset
8 publications focused on the age of UC patients with CMV infection or reactivation. The heterogeneity test of the included studies demonstrated significant heterogeneity (I2 = 64.6%, P = 0.006). The result of the random effect model showed that the older the age of onset of UC, the more likely to suffer from be CMV infection or reactivation. (MD = 6.212, 95% CI: 2.552 to 9.971, P = 0.001; Fig. 4).
Disease duration
9 articles described the relationship between the disease duration and the risk of CMV infection or reactivation. The heterogeneity test of the included articles indicated that there was heterogeneity among the studies. The results of random effect model analysis suggested that there was no statistical significance between the different course of UC patients and the risk of CMV infection or reactivation (heterogeneity P = 0.000, I2 = 72.4%, MD=-0.921, 95% CI: -2.423 to 0.582, P = 0.23; Fig. 5).
Glucocorticoid
14 articles described the risk of CMV infection or reactivation in UC patients treated with glucocorticoid. The heterogeneity test indicated that there was no heterogeneity (I2 = 38.1%, P = 0.073). The results of fixed analysis indicated that glucocorticoid treatment could significantly increase the risk of CMV infection or reactivation in UC patients, which was 4.175 times higher than that without glucocorticoid treatment (OR = 4.175, 95% CI: 3.076 to 5.666, P = 0.001; Fig. 6).
Immunosuppressant
15 articles described the risk of CMV infection or reactivation following immunosuppressive agents (azathioprine, cyclosporine A, methotrexate) in UC patients. Heterogeneity testing of the included literature indicated that there is heterogeneity among the literatures (I2 = 65.9%, P = 0.000). The results of the random effect model analysis showed that the use of immunosuppressive therapy would increase the risk of CMV infection or reactivation in patients in UC patients (OR = 2.038, 95% CI: 1.259 to 3.301, P = 0.004; Fig. 7).
Azathioprine
5 included articles reported on the use of azathioprine in UC patients. Heterogeneity test of included studies suggested no obvious heterogeneity (I2 = 0.0%, P = 0.501). There is no statistical difference on the Azathioprine (OR = 1.357, 95% CI: 0.831 to 2.217, P = 0.222; Fig. 8).
Infliximab
7 studies concentrated on whether the infliximab had side-effects on CMV infection or reactivation in UC patients. There was no obvious heterogeneity (I2 = 68.3%, P = 0.004); therefore, a random effect model was applied. Pooling the results demonstrated that infliximab treatment was not a risk factor for CMV infection or reactivation (OR = 1.915, 95% CI: 0.870 to 4.217, P = 0.107; Fig. 9).
5-ASA
9 articles described the risk of CMV infection or reactivation in UC patients treated with5-ASA. Interestingly, the results of the fixed-effect model showed that 5-ASA could reduce the risk of CMV infection or reactivation (heterogeneity P = 0.283, I2 = 18.0%, OR = 0.674, 95% CI: 0.492 to 0.924, P = 0.014; Fig. 10).
Sensitivity Analysis
Results of sensitivity analysis demonstrated that the effects of all the outcomes (severe UC, pancolitis, age of onset, disease duration, glucocorticoid, immunosuppressant, azathioprine, infliximab, 5-ASA) on CMV infection or reactivation in UC patients remained consistent after removing the trials one by one, as shown in Fig. 11.