To the Editor:
McIsaac et al.1 recently published their population-based cohort study on outcomes after total knee arthroplasty in relation to the use of peripheral nerve blocks. The primary outcome was length of stay (LOS), and they concluded that nerve blocks reduced LOS (risk ratio = 0.98!).
Although such large cohort studies may be valuable, we find the discussion insufficient in relation to the primary outcome, where we get no information on why the patients were hospitalized or whether a type of fast-track care was implemented.2 Furthermore, there is no information about discharge destination, which we know from several studies may hinder sufficient interpretation of LOS, because transfer of patients to rehabilitation or other institutions may depend on potential economic benefit3 or on local traditions4 and may misleadingly reduce the registered LOS after surgery.3 Finally, their mean LOS was approximately 4.7 days, which is beyond what has been published before (but not referred to) from prospective multicenter studies with a mean LOS of 3.0 days5 from well-defined fast-track programs without the use of peripheral blockades. Also, median values of LOS of approximately two days in subsequent large cohorts are available.6
In summary, when discussing LOS as a primary outcome, interventional studies in perioperative medicine need to include data on why the patient was hospitalized, as well as discharge destination.4
The authors declare no competing interests.