Time shift in early postoperative mortality after oesophagectomy for cancer.

BACKGROUND

Postoperative mortality is traditionally defined as death within 30 days of surgery. We hypothesised that the declining 30-day mortality after oesophageal cancer resection is, at least partly, explained by a shift towards increased 90-day mortality.

METHODS

This population-based cohort study included 95 % of all patients who underwent surgical resection for oesophageal cancer in Sweden in 1987-2010. Cox proportional-hazards regression models were used to calculate hazard ratios (HRs) with 95 % confidence intervals (CIs) of 30-day and 31-90 days postoperative mortality in three calendar periods (1987-1994, 1995-2002, and 2003-2010). Adjustments were made for age, sex, comorbidity, tumour stage, tumour histology, surgical radicality, neoadjuvant therapy, and hospital volume of oesophagectomy.

RESULTS

Among 1,822 patients, the 30-day postoperative mortality decreased from 9.3 % in 1987-1994 to 3.0 % in 2003-2010, while the corresponding 31-90 days mortality decreased from 8.4 to 4.6 %. The adjusted HR of 30-day mortality in the earliest period was markedly increased compared to the latest period (HR 3.26; 95 % CI 1.96-5.45), whereas the corresponding HR of 31-90 days mortality was weaker (HR 2.16; 95 % CI 1.34-3.46). Among patients who died within 90 days of surgery, the proportion of 31-90 days mortality increased from 47 to 61 % during the study period.

CONCLUSIONS

This population-based study indicates a shift of postoperative mortality following surgery for oesophageal cancer from 30 days to 31-90 days with more recent calendar periods. Reporting of 90-day mortality rates might replace 30-day mortality rates in assessing early postoperative mortality in oesophageal cancer patients.