Effect of COVID-19 on colorectal cancer care in England

During the rise of the COVID-19 pandemic in spring, 2020, unprecedented pressure on hospital beds and intensive care units (ICUs), redeployment of staff, caution regarding nosocomial transmission, reduced primary care access, and population lockdown combined in a perfect storm, dramatically disrupting UK cancer care pathways.¹ It was predicted that colorectal cancer care might fare particularly badly, in particular due to suspension of non-emergency diagnostic endoscopy following safety concerns from the British Society of Gastroenterology, discontinuation of the National Bowel Cancer Screening programme, recommendation by the Royal College of Surgeons against laparoscopic procedures, and shortage of ICU capacity to support open bowel resections.²

Under normal circumstances, 32% of colorectal cancers in England are typically diagnosed through the rapid access 2-week wait (2WW) urgent symptomatic referral pathway. The routes to diagnosis for the remaining colorectal cancers include emergency presentation (24%), screening (10%), and routine referral (34%), which includes those under long-term surveillance.³ With cessation of both screening and most routine outpatient activity, it was predicted that those colorectal cancer diagnoses would be displaced into the 2WW and emergency pathways.⁴

Using four population-based datasets spanning the National Health Service (NHS) in England, Eva Morris and colleagues present a comparison against the previous year's activity for the months of January to October, 2020, for colorectal 2WW presentations, colonoscopies, diagnoses, and treatment.⁵ They demonstrate peak reductions for April for 2WW referrals (63% reduction) and colonoscopies (92% reduction), with restitution to normal rates by October, 2020. The authors calculate there to have been a sustained relative reduction of 22% in the number of colorectal cancer cases referred for treatment across all routes to diagnosis from April to October, 2020. In total, they calculate that, across those 7 months, more than 3500 fewer people than in 2019 were diagnosed and treated for colorectal cancer in England.

Morris and colleagues offer the first clear quantitation of the drop in presentations, diagnosis, and treatment of colorectal cancer cases for England in 2020. An interesting question emerges regarding those apparently missing cancer diagnoses. Have they yet to appear as a downstream bulge of late or emergency presentations? Have some of them already been absorbed unnoticed within the COVID and non-COVID-related excess deaths of 2020? The data presented by Morris and colleagues do not include sex-specific or age-specific rates, which might provide additional insight into the demographic groups to which the missing cases correspond. Furthermore, comprehensive description of routes to diagnosis for this period once available will be informative.

What will be the ultimate impact in lives or life-years lost of this disruption? For any given patient presenting with seemingly localised cancer, surgery with curative intent can indeed cure them, restoring a near-normal life expectancy. Conversely, at the time of surgery, it may already be too late, and a seemingly localised tumour has already micrometastasised, with inevitable recurrence and premature death. Any delay to surgery will increase the likelihood of a patient moving from the first group to the second. Several groups have sought to quantify the impact per day, week, or month of treatment delay using linear regression from observational data to generate hazard rates that can be applied to routinely generated 5-year or 10-year stage-specific and age-specific survival data.6, 7, 8 However, although the reduction in activity shown by Morris and colleagues alludes to substantial disruption to colorectal cancer pathways, the actual extent of per-patient delay cannot be deduced. Delays in the 2WW pathway are available from the Cancer Waiting Times datasets, but these metrics fail to reflect delays in patient presentation, delays in primary care referral, or indeed delays in the other three routes to diagnosis. When cancer stage data become available for the 2020 colorectal cancer diagnoses, this will allow evaluation of net overall upwards stage-shifting, from which attributable excess colorectal cancer mortality can be indirectly predicted.⁴ However, only via analysis over the next decade for statistical deviation from expected colorectal cancer death rates can we attempt to quantify directly the excess mortality, as colorectal cancer deaths attributable to COVID-19-related disruption will be intermingled and indistinguishable from the expected colorectal cancer deaths within routinely reported statistics.

And what of the second wave of COVID-19? Urgent diagnostic colonoscopy has now been restored following consensus on appropriate measures for infection control. There has been aggressive public messaging around prompt presentation for symptomatic patients. Primary care is overall better prepared, with faecal immunochemical tests widely implemented for triage of symptomatic patients.⁹ However, while the clinical and health-economic cases for ring-fencing have been well-made, can cancer services really be protected in the face of acute pressure on capacity? It is politically challenging to prioritise the excess deaths of tomorrow over the emergencies of today.⁶ The data from Morris and colleagues well validate the long-stated case for dedicated stand-alone facilities of the sort elsewhere in Europe that have allowed continuity with minimal disruption of cancer diagnosis and treatment.¹⁰ We can only hope that the pandemic will prompt reconfiguration of cancer services to better protect future delivery in the face of the next extrinsic crisis.