Background Studies of mortality after cataract surgery have reported conflicting results: initial studies reported increases, while subsequent studies reported excess in the younger patients only. These studies compared mortality after cataract surgery with that after other operations, or with population rates. Neither comparison group is likely to be representative of patients having cataract surgery. Previously we have shown that plotting temporal profiles of mortality in successive periods after operation can show whether or not an operation is associated with excess postoperative mortality; and expressing mortality relative to that in the general population that is, as standardised mortality ratios (SMRs), can highlight the effects of selection as ‘fit for surgery’ and case-mix. We aimed to do this for cataract surgery.

Methods Records of all day cases and inpatient hospital admissions for cataract surgery in England, 1999–2010 linked to death records up to 2011 for patients aged 45 years and over. Death rates were calculated for consecutive months in year after operation and expressed as SMRs.

Results There were 2,538,835 operations followed by 4492 deaths (0.2%) within 30 days and 110,703 (4.4%) within one year of operation. The corresponding SMRs were 39 (95% confidence interval (CI) 37.5–39.8) and 78 (95% CI 77.8–78.8), both lower than the rates in the general population. SMRs were inversely related to age with the 45–54 age group having a one-year SMR of 333 while those aged 90+ years having a SMR of only 57. Plotting the SMRs for consecutive months after surgery showed that the first month had the lowest value, the values increased for the next two months and then stabilised for the rest of the year. This pattern was found for all the individual subgroups analysed – age-group, sex, with and without diabetes – though the actual values of the SMRs differed. The one month SMR was, on average, 45% that of the 4–12 month SMR.

Conclusion If cataract surgery was associated with excess postoperative mortality then a cluster of deaths occurring shortly after operation would be expected. No such cluster was found. The relatively low death rate in the first month after surgery is likely to be due to selecting patients as ‘fit for surgery’. The inverse relationship between SMR and age shows the effect of case-mix with the younger patients being relatively less healthy and the older patients being healthier than the general population.