Ethical approval or patient consent was not needed because this is a meta-analysis in which all data were extracted from published literature. We have performed a comprehensive computer bibliographic search on PubMed/MEDLINE, Google Scholar, EMBASE, Cochrane Library database, SciELO, Springer Link, Chinese Biomedical Database (CBD), China National Knowledge Infrastructure (CNKI) platforms, VIP, Chinese literature (Wan Fang) and China Science and Technology Journal database and Egyptian Knowledge Bank (EKB) Journals to identify all relevant studies concerning the proportion of the SARS-CoV-2 infection in patients with lung cancer published up to January31, 2021. We used the combination of the following search terms and keywords: (‘’Coronavirus Disease 2019’’ OR ‘’COVID-19’’ OR ‘’Severe Acute Respiratory Syndrome Coronavirus 2’’ OR ‘’SARS-CoV-2’’) AND (‘’chronic lymphocytic leukemia’’ OR ‘’CLL’’ OR ‘’multiple myeloma’’ OR ‘’MM’’ OR ‘’non-Hodgkin lymphoma’’ OR ‘’NHL’’ OR ‘’acute lymphoblastic leukemia’’ OR ‘’ALL’’ OR ‘’acute myeloid leukemia’’ OR ‘’AML’’ OR ‘’chronic myeloid leukemia’’ OR ‘’CML’’ OR ‘’Hodgkin lymphoma’’ OR ‘’HL’’) AND (‘’Pediatrics’’ OR ‘’Children’’ OR ‘’Adults’’ OR ‘’Female’’ OR ‘’Male’’). We restricted our search to pediatric and adult patients. The search was limited to English, Farsi and Chinese languages. Moreover, a manual search in all references of retrieved articles and reviews was carried out for more relevant articles.

The following inclusion criteria were used to select literatures for the meta-analysis: 1) consecutive case-series, case-control and cohort studies; 2) studies on COVID-19 prevalence in cancer patients with the SARS-CoV-2 infection, and; 3) sufficient data were presented to calculate the odds ratio (OR) with a 95% confidence interval (CI). The following exclusion criteria were used: 1) studies on only hematological cancers; 2) usable data not reported; 3) case only studies and non-consecutive case series; 4) in vitro studies; 5) studies with no available data and no contact with the authors; 6) posters, abstracts, letters, conference papers, non-standard data presentation, reviews and meta-analyses, and; 7) duplicate publications.

Two authors independently performed the publication search in the database. Subsequently, they reviewed the titles and abstracts of the selected studies in the primary search and extracted the necessary data into a form. When the authors were not in agreement, a third author was involved to reach an agreement. For each study, the following data ere extracted: first author's name, year of publication, country or region, ethnicity (the ethnicity was classified as Caucasian, Asian, African or mixed population), total numbers of infected cancer patients with the SARS-CoV-2 infection, age (mean and range), female/male ratio, number of hematological cancer patients with the SARS-CoV-2 infection and cancer treatment in total cases (surgical, radiotherapy, chemotherapy, targeted therapy and immunotherapy). If a duplicate publication or the same population (obvious overlap) was found in the primary literature search, the article with the larger sample size was selected for further analysis. The corresponding author for additional information, if the essential data was found to be incomplete, was contacted by email or telephone for any missing data.

Before the inclusion of selected studies in the meta-analysis, the methodological quality of the selected studies in the meta-analysis was performed using the Newcastle-Ottawa scale (NOS). The NOS ranges from zero to nine stars selection of patients (4 points), comparability of the groups (2 points) and ascertainment of exposure (3 points). Each selected study was interpreted to be of low quality (for scores ≤ 4), moderate quality (for scores 5 - 6) or high quality (for scores ≥ 7). Two authors assessed the quality of included studies independently and all disagreements were resolved by discussion or by consulting a third party.

The proportion of hematological cancer patients with the SARS-CoV-2 infection was assessed using odds ratios (ORs) with 95% confidence intervals (CIs) based on the COVID-19 infection between total cancer patients and hematological cancer patients. The Z-test was employed to assess the significance of pooled ORs, in which p < 0.05 was defined as the significance threshold. We utilized the I2 statistics (range of 0 to 100%; I2 = 0 - 25%, no heterogeneity; I2 = 25 - 50%, moderate heterogeneity; I2 = 50 –75%, large heterogeneity; I2 = 75 - 100%, extreme heterogeneity) and the χ2 based Q-statistic (p ≤ 0.10) to assess between-study heterogeneities. If between-study heterogeneity existed statistically, a random effect model (DerSimonian and Laird method) was adopted. Otherwise, a fixed-effect model (Mantel-Haenszel method) was used to combine the pooled data in the absence of heterogeneity. Stratified analysis was performed based on by ethnicity and country of origin to evaluate ethnic-specific results. A visual inspection of the funnel plot was used to assess whether our combined results could have been influenced by publication biases. Moreover, Egger's test was performed to assess the publication bias statistically, in which p < 0.05 was considered statistically significant. If the publication bias tests indicated bias existed, the Duval and Tweedie ‘‘trim and fill’’ method was used to adjust the bias. In the current meta-analysis, the Comprehensive Meta-Analysis (CMA) software version 2.0 (Biostat, USA) was used to combine the results of single studies and calculations of all the statistical analyses.

As shown in Figure 1, our initial search yielded 593 studies retrieved through the publication search. After removing duplicated studies, 341 studies were remained for further assessment. Subsequently, 268 studies were excluded after assessment of the titles and abstracts. In sequence, 59 studies were excluded because of not reporting useful data for the analysis, reviews, case reports, non-consecutive or case series, or were on only hematological cancers and we failed to contact their corresponding authors. Finally, a total of 14 studies 12,14,36–39,15,20,30–35 with a total 3,770 infected cancer patients and 685 hematological cancer cases with COVID-19 were ultimately found to be eligible for inclusion. The main characteristics of the selected studies are presented in Table 1. All of the included articles were written in English and published between 2020 and 2021. The sample size of all infected cancer patients ranged from 18 to 928 and 1 to 169 for hematological cancer patients. The study patients were from the China, France, United States, United Kingdom, Canada, Iran and Pakistan. Of them, eight studies were performed among Caucasians (with 2,848 all cancer patients and 544 hematological cancer patients) and six studies among Asians (with 922 all cancer patients and 141 hematological cancer patients). As shown in Table 1, quality scores ranged from 6 to 9, indicating that all included studies had high-quality scores.

Figure 1.

Flowchart of literature search and selection process in the meta-analysis.

(0.39MB).

The pooled data on the proportion of hematological cancer patients with the SARS-CoV-2 infection are shown in Table 2. The I2 and the Q-statistic tests showed that there was a significant heterogeneity. Thus, a random effect model (DerSimonian and Laird method) was used to assess the proportion of hematological cancer patients with the SARS-CoV-2 infection. When all eligible studies were included, the pooled data revealed that the proportion of infected hematological cancer patients during the COVID-19 pandemic was 16.5% (95% CI 0.130 - 0.208, p ≤ 0.001, Figure 2) globally. Moreover, a stratified analysis by ethnicity and country of origin was performed. The subgroup analysis by ethnicity showed that the proportion of hematological cancer patients with the SARS-CoV-2 infection was 18.8% (95% CI 0.154 - 0.227, p ≤ 0.001, Figure 3A) in Caucasian and 12.4% in Asian (95% CI 0.057 - 0.250, p ≤ 0.001, Figure 3B), respectively. Moreover, the subgroup analysis by country of origin revealed that the proportion of infected hematological cancer patients was highest in the United Kingdom (22.5%, 95% CI 0.175 - 0.285, Figure 4A), followed by France (17.1%, 95% CI 0.134 - 0.217, Figure 4B) and China (8.2%, 95% CI 0.062 - 0.109, Figure 4C), respectively.

Figure 2.

Forest plot for proportion of hematological cancer patients with SARS-CoV-2 infection during the COVID-19 pandemic in overall population.

(0.33MB).

Figure 3.

Forest plot for proportion of hematological cancer patients with SARS-CoV-2 infection during the COVID-19 pandemic in by ethnicity. A: Caucasian; and B: Asian.

(0.44MB).

Figure 4.

Forest plot for proportion of hematological cancer patients with SARS-CoV-2 infection during the COVID-19 pandemic in by country of origin. A: UK; B: France; and C: China.

(0.45MB).

In the current meta-analysis, there was statistically significant heterogeneity (I2 = 89.33%, p ≤ 0.001) in the global population (Table 2). Therefore, subgroup analyses by ethnicity (Asian and Caucasian) and country of origin (UK, France and China) were performed to evaluate the source of heterogeneity. The subgroup analyses revealed that the heterogeneity was not reduced, or did not disappear, by ethnicity. However, the heterogeneity was reduced in the Chinese (I2 = 35.95 and PH = 0.195) and was also dispersed in the French (I2 = 0.00 and PH = 0.877) hematological cancer patients with the SARS-CoV-2 infection, which indicated that the country of origin might be the source of heterogeneity in our pooled data.

It is defined that the sensitivity analysis is a repeat of a meta-analysis to substitute alternative ranges of values for results that were arbitrary or unclear. Thus, in the current meta-analysis, a sensitivity analysis, by omitting each individual study at any time, was performed to assess the influence of a single study on the pooled ORs by repeating the meta-analysis. The sensitivity analysis results revealed that our pooled data did not change via omitting each single study. The results were also achieved by excluding a study 34 that focused on European, North American (USA) and Asian (mixed population) hematological cancer patients with the SARS-CoV-2 infection, indicating that our combined data were statistically reliable.

The publication bias is the well-known major problem in a meta-analysis, which exists when the selected studies differ systematically from all studies that should have been selected. In this meta-analysis, both the Egger's test and Begg's test were performed to assess the potential publication bias in the literature. The shapes of the Begg's funnel plot and the Egger's test revealed that there was an evidence of publication bias in the overall population (PBegg's = 0.028; PEgger's = 0.141). Thus, the Duval and Tweedie ‘‘trim and fill’’ method was utilized to adjust the publication bias in the overall population in the hematological cancer patients with the SARS-CoV-2 infection. As shown in Figure 5, the funnel plot of the trim and fill method did not change visually, indicating that the combined pooled ORs are reliable.

Figure 5.

The funnel plots of publication bias for proportion of hematological cancer patients with SARS-CoV-2 infection during the COVID-19 pandemic in overall population. ‘’Blue’’ without and ‘’Red’’ with the Duval and Tweedie ‘‘trim and fill’’ method.

(0.11MB).