Global impact of the COVID-19 pandemic on cytopathology practice: Results from an international survey of laboratories in 23 countries
The last 3 authors contributed equally to this article.
Abstract
Background
To the authors' knowledge, the impact of the coronavirus disease 2019 (COVID-19) pandemic on cytopathology practices worldwide has not been investigated formally. In the current study, data from 41 respondents from 23 countries were reported.
Methods
Data regarding the activity of each cytopathology laboratory during 4 weeks of COVID-19 lockdown were collected and compared with those obtained during the corresponding period in 2019. The overall number and percentage of exfoliative and fine-needle aspiration cytology samples from each anatomic site were recorded. Differences in the malignancy and suspicious rates between the 2 periods were analyzed using a meta-analytical approach.
Results
Overall, the sample volume was lower compared with 2019 (104,319 samples vs 190,225 samples), with an average volume reduction of 45.3% (range, 0.1%-98.0%). The percentage of samples from the cervicovaginal tract, thyroid, and anorectal region was significantly reduced (P < .05). Conversely, the percentage of samples from the urinary tract, serous cavities, breast, lymph nodes, respiratory tract, salivary glands, central nervous system, gastrointestinal tract, pancreas, liver, and biliary tract increased (P < .05). An overall increase of 5.56% (95% CI, 3.77%-7.35%) in the malignancy rate in nongynecological samples during the COVID-19 pandemic was observed. When the suspicious category was included, the overall increase was 6.95% (95% CI, 4.63%-9.27%).
Conclusions
The COVID-19 pandemic resulted in a drastic reduction in the total number of cytology specimens regardless of anatomic site or specimen type. The rate of malignancy increased, reflecting the prioritization of patients with cancer who were considered to be at high risk. Prospective monitoring of the effect of delays in access to health services during the lockdown period is warranted.
Introduction
During the recent coronavirus disease (COVID-19) pandemic outbreak caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2),1 several national health organizations and different pathology scientific societies recommended reductions in routine health maintenance due to the health emergency, which also affected cytopathological practices around the world.2-5 Normally, cytological specimens are obtained routinely regardless of whether the procedure is a screening procedure for more invasive histopathological examinations or a complete diagnostic, prognostic, and predictive evaluation. However, during the COVID-19 pandemic, procedures leading to a cytological sample needed to be carefully evaluated with respect to the risks and benefits to the patient as well as the health care provider. Indeed, to maintain the efficiency of health systems and to reduce the risk of infection for patients and medical staff, screening procedures were minimized or postponed until the “flattening of the curve” could be accomplished.3, 4 However, cytopathologists still were asked to ensure timely malignancy-related diagnoses because any delay could lead to an increase in cancer-related mortality.6
Clinical recommendations may be difficult to apply in routine practice, and to our knowledge it is unclear how these guidelines were implemented. The perception of reduced cytological activity during the COVID-19 pandemic has not yet been studied via real-world, practice-based evidence generated from different laboratories worldwide. Currently, the only available data have been reported by single institutions, and demonstrate a decrease in cytological workload.7, 8 It is interesting to note that, despite the reduced activity, the rate of malignant diagnoses significantly increased.8 However, single-institution reports are not robust enough to draw reliable conclusions on a global scale or for assessment of the effect of the prioritization of cytological samples from patients considered to be at high risk of malignancy. Because the implementation of a nationwide network and registry of cytopathology diagnoses still is limited, worldwide data are difficult to gather without a collective effort. To fill this knowledge gap, a large number of cytopathologists from different countries reviewed their clinical reports to assess how cytological practices were impacted during the COVID-19 pandemic worldwide.
Materials and Methods
Survey
Data regarding the activity of cytopathology laboratories were collected through an international survey.
An Excel questionnaire template was distributed through email to members of the CytoESP Working Group (cytopathologists from the European Society of Pathology) (https://www.esp-pathology.org/working-groups/esp-working-groups/cytopathology.html) and to cytopathologists who have taken part in 1 of the 9 Annual National Molecular Cytopathology meetings in Naples, Italy (https://www.molecularcytopathology.com/), accounting for a total of 65 invited participants. Only a single email and no reminders were sent. Participants were asked to provide data regarding their cytopathology practice during the first 4 weeks of the COVID-19 national lockdown.
The study period was individualized for each institution due to the variability of the lockdown among countries. In countries in which lockdown did not take place, cytopathologists were asked to provide data from the first 4 weeks of the peak infection spread. To assess changes, if any, in cytopathological practice, participants were asked to provide the same data compared with the corresponding period in 2019. Questions included in the survey are listed in Figure 1. Specifically, participants were asked to report on the total number of processed cytological samples, the total number of exfoliative samples specifying the number of different specimen types, the total number of fine-needle aspiration (FNA) samples specifying the different sampling sites, and the distribution of diagnostic classes in nongynecological samples (nondiagnostic, negative, atypical, suspicious, and malignant).
Statistical Analyses
All analyses were performed using the R statistical platform (version 4.0.2).
Differences between the 2 periods with respect to the ratio of exfoliative-to-FNA samples and to the malignancy rate were summarized using a meta-analytical approach, treating each institution as a different study. In the first analysis, the odds ratio (OR) was used as summary measure. The random effects model of DerSimonian and Laird was a priori selected due to the anticipated heterogeneity among institutions. Statistical heterogeneity between institutions was assessed using the I2 statistic (ie, the percentage of total variability across institutions not due to sampling error). Standard thresholds were considered for the determination of I2: ≤25% for low heterogeneity, 26% to 50% for moderate heterogeneity, and >50% for high heterogeneity. Results were shown using forest plots.
Global differences between the 2 periods with respect to the percentage of samples for each single anatomic site were assessed using the Fisher exact test and the corresponding P values were adjusted for multiplicity using the Benjamini-Hochberg correction procedure.
The ratio between exfoliative and FNA specimens and the sample site distribution were evaluated taking into account results from respondents who provided both exfoliative and FNA data (39 respondents). The sample site list was built considering sites provided from all participants; sites accounting for <30 samples in the reference period (2019) and sites reported as “other” all were grouped in the “other sites” category.
Results
A total of 41 of 65 respondents (63%) from 23 countries worldwide (Azerbaijan [1 respondent], Belgium [1 respondent], Brazil [1 respondent], Croatia [1 respondent], Finland [1 respondent], France [2 respondents], Germany [1 respondent], India [1 respondent], Italy [7 respondents], Japan [1 respondent], Moldova [1 respondent], the Netherlands [1 respondent], Poland [1 respondent], Portugal [1 respondent], Slovenia [1 respondent], South Africa [1 respondent], Spain [3 respondents], Sweden [1 respondent], Switzerland [3 respondents], Turkey [2 respondents], Ukraine [1 respondent], the United Kingdom [2 respondents], and the United States [6 respondents]) joined the survey (Fig. 2). For the most part, data reflected single-institution activity (39 of 41 respondents; 95.1%), except in 2 instances (4.9%) in which multi-institutional data were provided, namely from the Pathological National Automated Archive (PALGA) Public Pathology Database of the Netherlands (https://www.palga.nl/en/public-pathology-database/) and from the National Health Laboratory Service of South Africa. Since the timing of COVID-19 lockdown differed among countries, as reported in Figure 2, each institution selected a 4-week time frame between March 1 and April 30, 2020, as the most significant health emergency period.
A total of 36 of the 41 respondents (87.8%) completed all required fields; in 2 instances (4.9%) only data relative to gynecological samples (Papanicolaou tests) were provided. With regard to the distribution of diagnostic classes, in 1 case (2.4%) suspicious and malignant diagnoses were merged; in another, only malignant diagnoses were reported; and, finally, in 1 case data were not reported.
Overall, data relative to 294,544 cytological samples, including 104,319 cytological specimens from the COVID-19 pandemic period and 190,225 cytological samples from the corresponding period in 2019, were provided, with an overall workload reduction of 45.3% (range, 0.1%-98.0%). Data for each single respondent are reported in Table 1. Data were anonymized and a number randomly was assigned to each respondent.
Respondent | COVID-19 Pandemic | Corresponding Period in 2019 | Difference, % |
---|---|---|---|
1 | 440 | 1361 | −67.7 |
2 | 164 | 622 | −73.6 |
3 | 547 | 1001 | −45.4 |
4 | 185 | 983 | −81.2 |
5 | 1207 | 3402 | −64.5 |
6 | 173 | 858 | −79.8 |
7 | 50 | 475 | −89.5 |
8 | 456 | 736 | −38.0 |
9 | 273 | 1329 | −79.5 |
10 | 436 | 2576 | −83.1 |
11 | 292 | 2532 | −88.5 |
12 | 289 | 495 | −41.6 |
13 | 704 | 2856 | −75.4 |
14 | 102 | 295 | −65.4 |
15 | 120 | 143 | −16.1 |
16 | 702 | 2784 | −74.8 |
17 | 858 | 2366 | −63.7 |
18 | 384 | 677 | −43.3 |
19 | 526 | 3099 | −83.0 |
20 | 87 | 353 | −75.4 |
21 | 94 | 615 | −84.7 |
22 | 2352 | 6214 | −62.1 |
23 | 398 | 1237 | −67.8 |
24 | 39 | 128 | −69.5 |
25 | 162 | 806 | −79.9 |
26 | 212 | 661 | −67.9 |
27 | 190 | 962 | −80.2 |
28 | 400 | 595 | −32.8 |
29 | 248 | 828 | −70.0 |
30 | 1615 | 1616 | −0.1 |
31 | 3531 | 8658 | −59.2 |
32 | 4 | 1783 | −99.8 |
33 | 3479 | 12,680 | −72.6 |
34 | 13,345 | 38,824 | −65.6 |
35 | 456 | 508 | −10.2 |
36 | 263 | 635 | −58.6 |
37 | 126 | 1191 | −89.4 |
38 | 10 | 500 | −98.0 |
39 | 198 | 697 | −71.6 |
40 | 68,429 | 79,116 | −13.5 |
41 | 773 | 3361 | −77.0 |
Total | 104,319 | 190,558 | −45.3 |
- Abbreviations: COVID-19, coronavirus disease 2019.
Because changes in cytological practice could modify the ratio between exfoliative versus FNA samples, a detailed analysis was performed; although the pooled analysis did not demonstrate a significant variation in the ratio of exfoliative to FNA samples between the COVID-19 pandemic and the reference period (OR, 0.89; 95% CI, 0.74-1.08), a very high heterogeneity among the institutions was observed (I2 of 95%) (Fig. 3).
For any single anatomic site, an absolute reduction in the total number of cases was observed consistently; this reduction was more evident (>50%) in samples from the cervicovaginal tract, urinary tract, breast, thyroid, salivary gland, soft tissue, anorectal region, and bone marrow, whereas it was less pronounced (<50%) in samples from 8 sites (serous cavities, lymph nodes, respiratory tract, central nervous system, gastrointestinal tract, pancreas, liver, and biliary tract); data are summarized in Table 2. Considering the contribution of any single anatomic site to the overall activity, significant decreases were observed in samples from the cervicovaginal tract, thyroid, anorectal region, and bone marrow (P < .05) during the COVID-19 pandemic compared with the corresponding period in 2019. Conversely, the percentage of samples from the urinary tract, serous cavities, breast, lymph nodes, respiratory tract, salivary gland, central nervous system, gastrointestinal tract, pancreas, liver, and biliary tract increased (P < .05). No significant variation in the percentage of soft-tissue samples was observed (Table 2).
Site | Overall No. | Percentage | ||||
---|---|---|---|---|---|---|
COVID-19 Pandemic | Corresponding Period in 2019 | Difference, % | COVID-19 Pandemic, % | Corresponding Period in 2019, % | Adjusted P | |
Cervicovaginal tract | 19,269 | 75,884 | −74.6 | 53.7 | 68.61 | <.001 |
Urinary tract | 3778 | 8379 | −54.9 | 10.53 | 7.58 | <.001 |
Serous cavities | 3101 | 4626 | −33.0 | 8.64 | 4.18 | <.001 |
Breast | 980 | 2248 | −56.4 | 2.73 | 2.03 | <.001 |
Lymph node | 2850 | 4651 | −38.7 | 7.94 | 4.2 | <.001 |
Thyroid | 1169 | 5551 | −78.9 | 3.26 | 5.02 | <.001 |
Respiratory tract | 2308 | 4606 | −49.9 | 6.43 | 4.16 | <.001 |
Exfoliative samples (n = 1892) (82%) | Exfoliative samples (n = 4007) (87%) | |||||
FNA sample (n = 416) (18%) | FNA sample (n = 599) (13%) | |||||
Salivary gland | 195 | 482 | −59.5 | 0.54 | 0.44 | .021 |
Soft tissue | 143 | 386 | −63.0 | 0.4 | 0.35 | .172 |
CNS | 901 | 1309 | −31.2 | 2.51 | 1.18 | <.001 |
Gastrointestinal tract | 81 | 161 | −49.7 | 0.23 | 0.15 | .005 |
Pancreas | 378 | 518 | −27.0 | 1.05 | 0.47 | <.001 |
Liver | 98 | 158 | −38.0 | 0.27 | 0.14 | <.001 |
Biliary tract | 54 | 94 | −42.6 | 0.15 | 0.08 | .004 |
Anorectal region | 6 | 183 | −96.7 | 0.02 | 0.17 | <.001 |
Bone marrow | 41 | 220 | −81.4 | 0.1 | 0.2 | .003 |
Other sites | 528 | 1153 | −54.2 | 1.5 | 1.04 | <.001 |
Total | 35,880 | 110,609 | 100% | 100% |
- Abbreviations: CNS, central nervous system; COVID-19, coronavirus disease 2019; FNA, fine-needle aspiration.
With regard to the distribution of diagnostic classes in nongynecological samples, an overall increase of 5.56% (95% CI, 3.77%-7.35%) in the malignancy rate during the COVID-19 pandemic compared with the corresponding period in 2019 was observed (Fig. 4). When the suspicious category also was added to the malignant category, the overall increase was 6.95% (95% CI, 4.63%-9.27%) (Fig. 5). The heterogeneity among institutions was found to be very high in both analyses (I2 of 81% and I2 of 87%, respectively).
Discussion
To our knowledge to date, the effect of COVID-19 on cytological practice has been documented by only a few reports reflecting single-institution experience.7, 8 The results of the current study demonstrated that the COVID-19 pandemic impacted cytology practices around the world by dramatically reducing the cytological specimen volume across specimen types by 45.3%. There are a number of explanations. First, screening programs were suspended or widely reduced according to recommendations issued by pathology scientific societies.3, 4 During the COVID-19 pandemic, the number of cervicovaginal tract samples obtained was dramatically reduced, both in absolute terms (−74.6%) and in proportion to the overall cytological sample volume (53.7% vs 68.61%; P < .001). Because it was recommended that cervical cancer screening activities be postponed rather than cancelled, future investigations once the health emergency is over are warranted to assess to what extent women have returned to cervical cancer screening programs.
Compared with the reduction in pap smears, the percentage of other exfoliative specimens, such as serous fluid, urine, and cerebrospinal fluid, demonstrated a significant increase, thereby explaining why there was no significant difference noted with regard to exfoliative versus FNA samples during the COVID-19 pandemic compared with the corresponding period in 2019.
A second explanation for the reduction in cytological samples lies in the fact that FNA specimens were limited to patients in whom a diagnosis rendered by the cytopathologist would immediately affect management. As an example, the current survey demonstrated an overall reduction in thyroid FNA samples both in absolute number (−78.9%) and, considering the overall cytological sample volume, in percentage (3.26% vs 5.02%; P < .001). It is interesting to note that the majority of asymptomatic thyroid nodules are not medically urgent9, 10; in addition, most differentiated thyroid cancers have an indolent clinical course, thus explaining the recommendation to postpone thyroid FNAs to the end of the health emergency, taking into account that a long and undefined “waiting time” between an endocrinologist's referral and performance of FNA generates anxiety for patients. Ideally, the decision to postpone the FNA should be taken by a multidisciplinary board, based on nodule location, ultrasound features, and clinical pathology laboratory data, in particular serum thyrotropin and calcitonin levels.9, 11 Dedicated guidelines, also addressing medicolegal issues, could be useful to assist the interventional cytopathologist in deciding to delay a thyroid FNA.
Compared with the reduction in thyroid FNAs, the overall reduction in cytology volume was less evident for specimens for which the rate of malignancy usually is higher. In fact, lymph node, respiratory tract, breast, and salivary gland specimens were reduced in absolute terms but, considering the overall cytological sample volume, their percentage was significantly increased compared with 2019. As an example, respiratory tract cytological specimens demonstrated a reduction of 49.9% but a relative increase in percentage compared with 2019 (6.43% vs 4.16%; P < .001). Moreover, respiratory tract FNA samples showed a slight relative increase (18% vs 13% in 2019) with respect to exfoliative specimens (82% vs 87% in 2019) (Table 2), suggesting a prioritization of FNA procedures that directly sample a suspicious lesion. The data from the current study indicate that, despite biosafety issues,12-14 which are especially relevant in the handling of specimens from the upper and lower airways, lung cytopathology still was relatively robust during the health emergency. A more focused survey could shed light on how cytopathologists applied the recommendations to limit the practice of rapid on-site evaluation to avoid smears air-drying before Romanowsky staining15 and to modify the alcohol content of liquid-based cytology collection medium.16
The overall data from the current study demonstrated a remarkable reduction in cytological workload across laboratory practices around the world, and also indicated that patients at high oncological risk were prioritized. The results also demonstrated an overall increase in the relative malignancy rate among nongynecological samples during the COVID-19 pandemic compared with the corresponding period in 2019 (+5.56%). This is even more evident when the suspicious category also is taken into account (+6.95%), which is conceivable considering that the risk of malignancy of this category is not negligible, generally ranging from 50% to 100%.17 The significance of this “relative” increase in the percentage of malignant and suspicious diagnoses should be investigated further, bearing in mind that the majority of laboratories had a dramatic decrease in the “absolute” volumes of many specimens.
Although the current study has provided robust data reflecting an international collective effort, several limitations should be highlighted. First, the study period was limited to 4 weeks between March 1 and April 30, 2020. This period of time does not necessarily correspond to the peak of the COVID-19 pandemic in countries still facing the health emergency at the time of this writing (eg, Brazil, India, and South Africa). Second, the data were analyzed globally, which may conceal differences among institutional practices. In this setting, further investigations may be warranted when also taking into consideration that certain continents are better represented than others and additional data from Asia and Oceania would make the survey more complete.
Conclusions
The COVID-19 pandemic dramatically impacted health systems and the activity of cytopathology laboratories worldwide. Laboratories universally experienced a dramatic reduction in overall cytological specimen volume across specimen types, which resulted in a higher malignancy rate overall. Although the increase in the percentage of malignant cases demonstrates the efficacy of prioritizing high-risk patients with cancer despite the pandemic, prospective monitoring of the effect of delays in access to health services during the COVID-19 pandemic warrants further investigation.
Funding Support
Supported by Monitoraggio ambientale, studio ed approfondimento della salute della popolazione residente in aree a rischio–In attuazione della D.G.R. Campania n.180/2019; POR Campania FESR 2014-2020 Progetto “Sviluppo di Approcci Terapeutici Innovativi per patologie Neoplastiche resistenti ai trattamenti (SATIN)”; and Campania Region for the Investigation of the Molecular Biology of Thyroid Cancer (grant LR n.24 29/12/2005).
Conflict of Interest Disclosures
Umberto Malapelle has received personal fees for acting as a member of the Speakers' Bureau or as an advisor from Boehringer Ingelheim, AstraZeneca, Roche, MSD, Amgen, Merck, Eli Lilly, and Diaceutics for work performed outside of the current study. Paul A. VanderLaan has acted as a paid consultant for Gala Therapeutics, Foundation Medicine, Caris Life Sciences, Flatiron Health, Intuitive Surgical, and Clearview Healthcare Partners for work performed outside of the current study. Giancarlo Troncone has received personal fees as a member of the Speakers' Bureau or as an advisor from Roche, MSD, Pfizer, and Bayer for work performed outside of the current study. The other authors made no disclosures.
Author Contributions
Conceptualization: Elena Vigliar, William C. Faquin, Martha Bishop Pitman, and Giancarlo Troncone. Methodology: All authors. Software and formal analysis: Elena Vigliar and Dario Bruzzese. Validation: All authors. Investigation: All authors. Resources: All authors. Data curation: All authors. Writing–original draft: Elena Vigliar and Giancarlo Troncone. Writing–review and editing: All authors. Visualization: All authors. Supervision: Elena Vigliar, William C. Faquin, Martha Bishop Pitman, and Giancarlo Troncone. Project administration: Elena Vigliar, William C. Faquin, Martha Bishop Pitman, and Giancarlo Troncone. Funding acquisition: Giancarlo Troncone.