Burning the midnight oil does not impact surgical outcomes

Key clinical point: The risks of adverse outcomes of elective daytime surgical procedures were similar whether or not the attending physician had provided medical care the previous night.

Major finding: The primary composite endpoint was similar between the postmidnight and control groups (22.2% vs. 22.4%; P = .66).

Data source: Population-based, retrospective, matched cohort study in 38,978 patients.

Disclosures: Dr. Govindarajan reported grant support from the Canadian Institutes for Health Research and the University of Toronto Dean’s Fund while conducting the study.




Attending physicians who work through the wee hours of the night do not have measurably different short-term outcomes for elective surgeries performed the next day, according to a population-based, matched-control study published online Aug. 26.

The primary composite outcome of death, readmission, or complications within 30 days occurred in 22.2% of patients undergoing elective daytime surgery by an attending who treated patients from midnight to 7 a.m. and in 22.4% of those undergoing the same procedure by the same attending, but after a night when no clinical work had been performed (P = .66).


There was no significant between-group difference in the primary outcome in adjusted analyses (adjusted odds ratio, 0.99; P = .65).

Secondary outcomes also were similar between the postmidnight and control groups: death within 30 days (both 1.1%), readmission within 30 days (6.6% vs. 7.1%), complications within 30 days (18.1% vs. 18.2%), median length of stay (both 3 days), and median duration of surgery (both 2.6 hours).

“These data suggest that calls for broad-based policy shifts in duty hours and practices of attending surgeons may not be necessary at this time,” wrote surgical oncologist Dr. Anand Govindarajan of Mount Sinai Hospital, Toronto, and his associates (N Engl J Med. 2015;373:845-53. doi: 10.1056/NEJMsa1415994).

The current study addresses a gap in the literature on the effects of sleep deprivation and may help inform policy discussions on this issue, the authors suggested.

Most studies of physicians suggesting that sleep deprivation may affect mood, cognition, and psycho-motor function have focused on medical trainees, but few studies have examined the effects of sleep deprivation in attending physicians and the results have been conflicting.

A 2009 single-center study prompted calls for policy-level changes regarding sleep deprivation in surgeons after showing a higher rate of complications for procedures performed by attending physicians with sleep opportunities of less than 6 hours (JAMA. 2009 Oct 14;302[14]:1565-72), but the findings have not been replicated by other groups, Dr. Govindarajan and his associates observed.

In the current study, a small but significant increase in complications was observed in the subset of patients whose physician had performed two or more procedures the night before (adjusted OR, 1.14; P = .05). Importantly, this isolated finding was from a post-hoc subgroup analysis and may be the result of random error, they suggested.

Further, three a priori analyses found no significant difference in outcomes after stratification for academic vs. nonacademic center, physician age, or type of procedure.

The current study involved 38,978 patients who underwent 1 of 12 elective daytime procedures performed by 1,448 physicians at 147 hospitals in Ontario. Patients undergoing procedures performed by a physician who had treated patients from midnight to 7 a.m. were matched in a 1:1 ratio to patients undergoing the same procedure by the same physician on a day when the physician had not treated patients after midnight.

The physicians had been in practice for a median of 20 years, and 40.6% of procedures were performed at academic institutions. Physicians who treated patients after midnight performed a mean of 1.25 procedures during that time.

The elective procedures (cholecystectomy, gastric bypass, colon resection, coronary artery bypass grafting, coronary angioplasty, knee replacement, hip replacement, hip fracture repair, hysterectomy, spinal surgery, craniotomy, and lung resection) were all performed on weekdays.

The investigators selected procedures that varied in duration and were associated with a range of complication rates because sleep studies have suggested that tasks requiring longer periods of concentration may be more affected by sleep deprivation.

“The broad scope of the study enhances its generalizability, a particularly relevant consideration if policy changes are being contemplated with respect to duty hours,” Dr. Govindarajan and his associates noted.

Other strengths of the study are the matched study groups and sufficiently sized cohorts and event rates to provide adequate power to show clinically meaningful differences.

Limitations of the study are the inability to quantify the number of hours that a physician was deprived of sleep and to determine whether there was a difference in outcomes between daytime procedures performed later in the day and those performed earlier in the day or whether procedures may have been postponed till later in the day because of substantial sleep deprivation. “However, given the constraints of operating room schedules in Ontario, it would not usually be feasible to postpone an operation until later in the day on short notice,” the investigators wrote.

Dr. Govindarajan reported grant support from the Canadian Institutes for Health Research and the University of Toronto Dean’s Fund while conducting the study.

Next Article:

   Comments ()