Guest blog: Death following pulmonary complications of surgery before and during the SARS-CoV-2 pandemic: how to compare apples and oranges?

Early in the pandemic, it was recognised that patients had a higher rate of death associated with pulmonary complications if they become infected with SARS-CoV-2 in the perioperative period ¹. However, the key question is why? Was this solely down to SARS-CoV-2 infections, was this due to differences in who was prioritised for surgery, or due to some other reason? These are difficult questions to answer, especially without comparable data from the pre-pandemic period. 

To address this, our paper published in the BJS pools together two large-scale, prospective and multi-specialty international cohorts ²: one from the pre-pandemic period from the STARSurg Collaborative ³ and another from during the pandemic from the COVIDSurg collaborative ⁴. Since these projects both included elective abdominal cancer surgery patients, and collected similar data on them and their postoperative outcomes, we were able to make a fair comparison between these cohorts. 

Overall, there was only a minority of patients with a perioperative SARS-COV-2 infection (4.3%), which is reassuring that most patientsundergoing surgery during this period were safeguarded. Interestingly, when postoperative outcomes were compared between the periods, the pulmonary complication rate observed was relatively similar (7.1% [before] vs 6.3% [during], p=0.158) but the 30-day mortality was much higher (0.7% [before] versus 2.0% [during], p<0.001). This was perhaps due to more “severe” pulmonary complications during the pandemic, e.g. a higher proportion of patients with acute respiratory distress syndrome. 

Overall, we estimated the number of deaths related to pulmonary complications (the “population attributable fraction”) increased from 37.0% (95% CI: 14.6% – 64.1%) pre-pandemic to 66.0% (95% CI: 48.6% – 79.3%) during the pandemic. This is (perhaps unsurprisingly) difficult to tease out the exact role of SARS-COV-2 infection in this – not least because SARS-COV-2 infections didn’t exist in the pre-pandemic group. We also found some major differences in patient selection and surgical practice during the pandemic. For example, while patients operated during the pandemic tended to have an overall lower baseline risk of pulmonary complications, they still had higher rate of open operations (this may be at least in part due to uncertainty at the time whether or not laparoscopic surgery was high-risk for COVID-19 transmission ⁵). To allow us to account for these differences in the cohorts, we used a technique called “mediation analysis”, which allows us to estimate exactly how much SARS-COV-2 infection contributed to (“mediated”) 30-day postoperative mortality during the pandemic.

We found that even after adjustment, there was still an almost 3-times higher likelihood of death when being operated on during the pandemic, compared to pre-pandemic (OR: 2.72, 95% CI: 1.58 to 4.67, p<0.001). However, we estimated that over half of excess deaths (54.8%) during the pandemic were explained by the presence of a SARS-CoV-2 infection. If in a parallel universe these patients were operated on pre-pandemic (e.g. with no risk of SARS-CoV-2 infection) there would still have likely been a higher rate of death than expected (perhaps due to a collateral impact on other hospital services from the pandemic that reduced capacity to rescue). However, overall, there would have been no significant difference in postoperative mortality during the pandemic (OR:1.57, 95% CI: 0.91 to 2.73, p=0.108), compared to pre-pandemic. 

In the first pandemic wave, despite attempts by care providers to continue safe elective surgery by operating on lower-risk patients, a significantly higher risk-adjusted mortality was observed. Whilst over 50% of excess deaths were explained by SARS-CoV-2 infection, the remainder may indicate a collateral impact on other hospital services that reduced capacity to rescue. The COVID-19 pandemic isn’t over yet, and upscaling elective surgery safely to meet the backlog is one of the major challenges facing health systems going forward. This study highlights the ongoing need for governments to ensure hospitals have Covid secure pathways ⁶ and to take appropriate and timely measures to ensure hospitals services aren’t overwhelmed. Several million patients have had their cancer surgeries delayed or cancelled ⁷, and there are serious concerns that despite best efforts to prioritise care this may lead to reductions in cancer survival in the long-term ⁸. It is likely to be several years until universal vaccination is available, and so until then strategies to mitigate risk of SARS-COV-2 infection must be implemented to continue surgery safely.