You are here

Article Text

Article menu
Original research
Effects of night surgery on postoperative mortality and morbidity: a multicentre cohort study
  1. Friederike C Althoff1,
  2. Luca J Wachtendorf1,
  3. Paul Rostin1,2,
  4. Peter Santer1,
  5. Maximilian S Schaefer1,
  6. Xinling Xu1,
  7. Stephanie D Grabitz1,
  8. Hovig Chitilian2,
  9. Timothy T Houle2,
  10. Gabriel A Brat3,
  11. Oluwaseun Akeju2,
  12. Matthias Eikermann1
  1. 1 Anesthesia, Critical Care and Pain Medicine, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA
  2. 2 Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, Massachusetts, USA
  3. 3 Surgery, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA
  1. Correspondence to Professor Matthias Eikermann, Anesthesia, Critical Care and Pain Medicine, Beth Israel Deaconess Medical Center, Boston, MA 02215, USA; meikerma{at}bidmc.harvard.edu

Abstract

Background Surgery at night (incision time 17:00 to 07:00 hours) may lead to increased postoperative mortality and morbidity. Mechanisms explaining this association remain unclear.

Methods We conducted a multicentre retrospective cohort study of adult patients undergoing non-cardiac surgery with general anaesthesia at two major, competing tertiary care hospital networks. In primary analysis, we imputed missing data and determined whether exposure to night surgery affects 30-day mortality using a mixed-effects model with individual anaesthesia and surgical providers as random effects. Secondary outcomes were 30-day morbidity and the mediating effect of blood transfusion rates and provider handovers on the effect of night surgery on outcomes. We further tested for effect modification by surgical setting.

Results Among 350 235 participants in the primary imputed cohort, the mortality rate was 0.9% (n=2804/322 327) after day and 3.4% (n=940/27 908) after night surgery. Night surgery was associated with an increased risk of mortality (ORadj 1.26, 95% CI 1.15 to 1.38, p<0.001). In secondary analyses, night surgery was associated with increased morbidity (ORadj 1.41, 95% CI 1.33 to 1.48, p<0.001). The proportion of patients receiving intraoperative blood transfusion and anaesthesia handovers were higher during night-time, mediating 9.4% (95% CI 4.7% to 14.2%, p<0.001) of the effect of night surgery on 30-day mortality and 8.4% (95% CI 6.7% to 10.1%, p<0.001) of its effect on morbidity. The primary association was modified by the surgical setting (p-for-interaction<0.001), towards a greater effect in patients undergoing ambulatory/same-day surgery (ORadj 1.81, 95% CI 1.39 to 2.35) compared with inpatients (ORadj 1.17, 95% CI 1.02 to 1.34).

Conclusions Night surgery was associated with an increased risk of postoperative mortality and morbidity. The effect was independent of case acuity and was mediated by potentially preventable factors: higher blood transfusion rates and more frequent provider handovers.

  • anaesthesia
  • adverse events
  • epidemiology and detection
  • surgery
  • standards of care

Data availability statement

Data are available on reasonable request. Due to the sensitive nature of the data collected for this study, requests to access the dataset from qualified researchers trained in human subjects research and confidentiality may be sent to Matthias Eikermann at meikerma@bidmc.harvard.edu.

https://doi.org/10.1136/bmjqs-2020-011684

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    More than 5% of surgeries are performed outside of standard operating room hours and up to 74% of procedures performed at night are planned as elective cases.1 Although unavoidable in urgent circumstances, night surgery itself may be an independent risk factor of perioperative mortality and morbidity.

    Previous studies analysing the effect of night surgery on mortality and morbidity provided equivocal findings: while higher mortality rates were suggested in patients undergoing night surgery across surgical specialties,2 other studies only found surgery during the early (13:00 to 17:00 hours)3 or late afternoon and evening (16:00 to 23:00 hours)4 to be associated with higher mortality or morbidity in non-emergency cases. It remains unclear whether surgery at night affects patient survival, and whether the effect varies by acuity and condition of the patient.

    It is important to identify potentially preventable elements of night surgery that contribute to its potential risk. Discussions aiming at understanding mechanisms that may drive a higher risk of night surgery have only focused on providers’ sleep deprivation. Studies on healthcare providers suggested that night shift work and variation in daylight exposure disrupt circadian rhythms, leading to increased medical errors.5–8 Contributing factors may be sleep-deprived fatigue, prolonged reaction time, impaired situational awareness and a natural circadian variability in technical performance.5 6 9 However, it has not yet been defined whether there are intraoperative factors that notably differ between day and night surgeries and may be modified to improve outcomes.

    In this multicentre cohort study, we hypothesised that exposure to surgery at night affects patient survival and postoperative complications. We then tested whether intraoperative blood transfusions and anaesthesia handovers during night-time were potential mediators of the effect of night surgery on adverse outcomes. We further assessed whether the effect of night surgery was driven by a higher case acuity of patients undergoing surgery at night.

    Methods

    Study design

    Data were obtained for surgical patients at institution A between October 2005 and September 2017 and at institution B and one affiliated community hospital between January 2007 and December 2015. The study was approved by the institutional review boards at each institution (protocol numbers 2018P000786 and 2019P000825). The requirement for written informed consent was waived. Data were collected from hospital-registry databases (online supplemental file 1, section 1). This manuscript adheres to the STROBE guidelines online supplemental file 2.10

    Supplemental material

    Supplemental material

    Study population

    We included patients aged 18 years or older who had an American Society of Anaesthesiologists' (ASA) status below 6 and underwent non-cardiac surgery with general anaesthesia. Missing data required for the primary analysis were imputed using multiple imputation with chained equations (online supplemental file 1, section 3). For secondary, sensitivity, and exploratory analyses, the complete-case method was used.

    Study exposure and outcomes

    Night surgery was defined as a surgical incision time between 17:00 and 07:00 hours. Surgical times were based on timestamps in electronic anaesthesia records. The primary outcome was mortality within 30 days of surgery. Secondary outcomes were 30-day morbidity, defined as a composite outcome including renal, cardiovascular, bleeding, infection, intestinal/digestive, and pulmonary complications within 30 days of surgery, as defined by the International Classification of Diseases, Ninth and Tenth Revision (online supplemental eTable 1)11 and the mediating effect of blood transfusion rates and provider handovers on the effect of night surgery on outcomes.

    Primary analysis and covariates

    We used multivariable-adjusted mixed-effects logistic regression to investigate the effect of night surgery on 30-day mortality. We included individual anaesthesia and surgical providers as random effects in the primary model. Analyses were adjusted for confounding variables based on literature review and clinical plausibility (online supplemental eTable 2). Patient factors included age, sex, body mass index, ASA physical status classification, surgical care setting (ambulatory care setting and same-day discharge versus inpatient surgery), Charlson Comorbidity Index, a 1-year history of cancer, home oxygen or respirator dependency, chronic pulmonary disease, coronary artery disease, congestive heart failure and ischaemic stroke. Procedure-related confounding variables included emergency surgery, surgical specialty, duration of surgery, date of surgery and work relative value units. Anaesthesia-related factors were vasopressor dose, intraoperative hypotension (defined as mean arterial pressure <55 mm Hg), SpO2/FiO2-ratio, transfusion of packed red blood cell units (PRBC) and handover between anaesthesiologists. Finally, we included the institution (A versus B) in the model. To address potential bias due to missing data, we performed multiple imputation with chained equations for the primary analysis.

    Secondary analyses

    In secondary analyses, we investigated whether night surgery was associated with 30-day morbidity by using multivariable-adjusted logistic regression. Further, we used path mediation analysis to evaluate the role of intraoperative factors as potential mediators of the effect of night surgery on 30-day mortality and morbidity. The two mediator candidates analysed were the intraoperative transfusion rate (proportion of patients transfused with PRBC) and the proportion of cases with an intraoperative handover between anaesthesiologists. To define complete handovers between anaesthesiologists, we used unique provider identification numbers as well as sign-in and sign-out times.

    First, we tested the hypothesis that intraoperative transfusion rates were higher during night cases than day cases. We used a multivariable-adjusted logistic regression model on the association between night surgery and transfusion rate that included all confounding variables of the primary analysis. To verify this observation, we additionally adjusted for intraoperative estimated blood loss (mL) as well as mild and moderate to severe anaemia within 30 days prior to surgery. Preoperative haemoglobin values of ≥11 to ≤12.9 g/dL in men and ≥11 to ≤11.9 g/dL in women were used to define mild anaemia, and values of ≤10.9 g/dL were used to define moderate to severe anaemia in both men and women.12 In addition, we tested whether variability across individual surgeons and anaesthesiologists had an impact on the association between night surgery and transfusion rate by including provider-related confounding variables and random effects for individual providers into the model (online supplemental file 1, section 5.1). Similarly, we tested the hypothesis that handovers between anaesthesiologists occur more frequently during night-time.

    Second, we used adjusted analyses to examine whether transfusion rates and handovers were associated with the outcomes of 30-day mortality and morbidity, indicating potential effect mediation. We additionally tested for a potential interaction between the effects of the two mediators on the outcomes including the interaction term ‘transfusion rate*handover’ in the model.

    Conditional on an association between the mediators and the study outcomes, we performed adjusted formal mediation analysis based on the method by Buis.13 We estimated ORs of the indirect (mediated) effect of transfusion rate and handovers, respectively, and the total (unmediated) effect of night surgery on mortality and morbidity, using bootstrapping with 1000 replications.13–15 Percentage mediation by the mediators was calculated using the following form: (ln (indirect effect)/ln (total effect)) × 100.15 Finally, we combined the two mediators into one model to test the mediating effect of both higher proportions of blood transfusion and handovers during night-time on the association between night surgery and outcomes.

    Exploratory analyses

    With an exploratory intent, we performed interaction analysis to assess whether the association between night surgery and mortality was modified by the surgical care setting (ambulatory and same-day surgery versus inpatient surgery). Subgroup analyses were performed across groups of the interaction term. Patient characteristics in subgroups are provided in online supplemental eTables 6 and 7. To further address potential concerns that the effect may be driven by a higher case acuity of patients operated at night, we tested the primary association in a subgroup of patients undergoing non-emergency surgery, after excluding emergency and acute care services. In this subgroup, we repeated the interaction analysis by surgical care setting. Conversely, we examined the effect of night surgery on mortality in a subgroup of patients undergoing emergency surgery (online supplemental eTable 8).

    Sensitivity analyses

    We performed several sensitivity analyses to confirm the robustness of the effect of night surgery on mortality. We repeated the primary multivariable-adjusted mixed-effects logistic regression analysis with random effects for anaesthesia and surgical providers in the complete-case cohort after excluding cases with missing confounder data. We performed propensity score matching and provide patient characteristics in the propensity-matched cohort in online supplemental eTable 9. We further provide data on postoperative major adverse events within 30 days, indicating potential causes of death (online supplemental file 1, section 4). In addition, we tested whether the association between night surgery and mortality was modified by a recent diagnosis of cancer by including an interaction term between night surgery and 1 year diagnosis of cancer (online supplemental file 1, section 6.2). We tested for potential effect modification of the association between night surgery and mortality by case delays (ie, difference between scheduled and actual start time in minutes) by including an interaction term between night surgery and case delay. We also investigated the mediating effect of case delays on the association between night surgery and mortality (online supplemental file 1, section 5.2). Finally, we conducted several subgroup analyses to ensure that the impact of night surgery on mortality was not driven by a narrow patient population. Details on all sensitivity analyses are described in section 6 of online supplemental file 1.

    Sample size justification

    Assuming a two-sided alpha level of 0.05, an observed proportion of night surgeries of 8.0% in the primary cohort, a baseline mortality risk of 0.6%16 and a clinically significant risk increase of 30% in the group of patients who underwent night surgery, the sample size of this study provided a power of 94.0% to detect a difference between patients undergoing surgery during the night versus day.

    Statistical analyses

    Confounding variables demonstrating linear associations with the primary outcome were included as continuous, whereas variables with non-linear associations were categorised into quintiles (online supplemental file 1, section 2). A two-sided p<0.05 was considered statistically significant. Analyses were performed using Stata (Stata, V.13) and RStudio (RStudio, V.3.2.5).

    Patient and public involvement

    No patients or the public were involved in the design or conduct of this study. Patients or the public were not invited to contribute to the writing or editing of this document for readability or accuracy.

    Results

    Study cohort

    In this study, 377 163 cases were considered for inclusion. After application of exclusion criteria and imputation of missing data, the final cohort for the primary analysis consisted of 350 235 cases (figure 1). A total of 322 327 (92.0%) patients underwent day surgery (07:00 to 17:00 hours) and 27 908 (8.0%) underwent night surgery (17:01 to 06:59 hours). In the complete-case cohort, 303 892 cases remained after excluding cases with missing data. Patient characteristics and distribution of variables by study groups are detailed in table 1 and online supplemental eTables 3–5.

    Figure 1
    Figure 1

    Study flowchart. ASA, American Society of Anesthesiologist’s Physical Status Classification System; RVU, relative value unit.

    Figure 2
    Figure 2

    Path mediation analyses for 30-day mortality and morbidity. Mediation effects of (a) intraoperative blood transfusion rates, (b) handovers of anaesthesia care as well as (c) a combination of both mediators on the association between night surgery and 30-day mortality (A) and morbidity (B) are shown. Transfusion rates mediated 5.9% and 4.9% of the effect of night surgery on 30-day mortality and morbidity, respectively. Handovers mediated 4.1% of the effect of night surgery on 30-day morbidity. Combination of handovers and transfusion rates mediated 9.4% and 8.4% of the effect of night surgery on 30-day mortality and morbidity, respectively. Adjusted ORs are shown for total, indirect and direct effect.

    View this table:
    Table 1

    Characteristics and distribution of variables by day versus night surgery in primary cohort

    Primary analysis

    30-day mortality

    In total, 3744 (1.5%) patients died within 30 days after surgery, 2804 (0.9%) after day surgery and 940 (3.4%) after night surgery. Night surgery was associated with increased mortality within 30 days of surgery in unadjusted (OR 2.45, 95% CI 2.26 to 2.65, p<0.001) as well as adjusted analyses (adjusted OR (ORadj) 1.26, 95% CI 1.15 to 1.38, p<0.001; table 2).

    View this table:
    Table 2

    Results of 30-day mortality (primary outcome) and 30-day morbidity (secondary outcome) associated with night surgery

    Secondary analyses

    30-day morbidity

    In total, 2234 (0.7%) patients had complications within 30 days after surgery, 1769 (0.6%) after day surgery and 465 (2.3%) after night surgery. We found a significant association between night surgery and morbidity (ORadj 1.41, 95% CI 1.33 to 1.48, p<0.001; table 2).

    Transfusion rate

    The proportion of patients receiving intraoperative blood transfusions was 5.5% (n=1149) during night surgeries and 2.8% (n=7882) during day surgeries. Night surgery was significantly associated with a higher transfusion rate in adjusted analysis (ORadj 1.25, 95% CI 1.15 to 1.35, p<0.001). The effect remained robust in a subgroup of cases where data on intraoperative estimated blood loss and preoperative haemoglobin levels were available, after additional adjustment for blood loss, mild anaemia, and moderate to severe anaemia within 30 days prior to surgery (ORadj 1.17, 95% CI 1.04 to 1.31, p=0.009; n=1 31 371), as well as when additionally accounting for a potential variability in transfusion practice across individual anaesthesia providers (ORadj 1.19, 95% CI 1.10 to 1.30, p=0.005; n=1 28 691) and across individual surgeons (ORadj 1.35, 95% CI 1.21 to 1.51, p<0.001, n=1 62 907; online supplemental file 1, section 5).

    Higher transfusion rate was significantly associated with an increased risk of 30-day mortality (ORadj 1.72, 95% CI 1.50 to 1.98, p<0.001) and morbidity (ORadj 1.68, 95% CI 1.59 to 1.78, p<0.001). Higher transfusion rates during night surgery mediated 5.9% (95% CI 2.4% to 9.3%; p=0.001) of the effect of night surgery on mortality, and 4.9% (95% CI 3.8% to 6.1%; p<0.001) of its effect on morbidity (table 3 and figure 2).

    View this table:
    Table 3

    Adjusted ORs with 95% CIs and p values obtained from path mediation analysis of intraoperative blood transfusion rates and handovers of anaesthesia care as potential mediators in the association between night surgery and 30-day mortality and morbidity

    Handover of anaesthesia care

    Anaesthesia handovers occurred in 22.6% (n=4686) of night surgeries and 8.8% (n=24 949) of day surgeries. Night surgery was significantly associated with a higher proportion of handovers in adjusted analysis (ORadj 4.12, 95% CI 3.94 to 4.30, p<0.001). Anaesthesia handovers were significantly associated with an increased risk of morbidity (ORadj 1.10, 95% CI 1.05 to 1.14, p<0.001) and mediated 4.1% (95% CI 2.3% to 5.9%, p<0.001) of the effect of night surgery on morbidity in formal mediation analysis (table 3).

    There were no interaction effects between the two mediators on 30-day mortality (p-for-interaction: 0.395) and morbidity (p-for-interaction: 0.125). When combining both mediators into one mediation model, higher transfusion rates together with more handovers during night-time mediated 9.4% (95% CI 2.4% to 16.4%, p<0.001) of the effect of night surgery on 30-day mortality and 8.4% (95% CI 6.3% to 10.4%, p<0.001) of its effect on morbidity (table 3 and figure 2).

    Exploratory analyses

    Ambulatory and same-day surgery versus inpatient surgery

    The effect of night surgery was modified by the surgical care setting (p-for-interaction<0.001) towards a greater effect among patients who underwent ambulatory or same-day surgery (ORadj 1.81, 95% CI 1.39 to 2.35, p<0.001) compared with inpatient surgery (ORadj 1.17, 95% CI 1.02 to 1.34, p=0.026; table 4).

    View this table:
    Table 4

    Results of 30-day mortality across exploratory analyses

    Non-emergency surgery

    Among 282 526 patients undergoing non-emergency surgery, night surgery was associated with 1.35-fold higher adjusted odds for 30-day mortality compared with day surgery, respectively (95% CI 1.16 to 1.56, p<0.001; table 4). Analysis of this subgroup demonstrated robust results when including an interaction term by the surgical care setting (p-for-interaction=0.001; ambulatory/same-day surgery: ORadj 2.22, 95% CI 1.60 to 3.07, p<0.001; inpatient surgery: ORadj 1.20, 95% CI 1.01 to 1.43, p=0.036) (table 4).

    Emergency surgery

    Characteristics of emergency cases by night versus day surgery are provided in online supplemental eTable 8. Of 13 566 emergency surgeries, 52.8% (n=7162) were performed at night and 47.2% (n=6404) were performed during the day. Distributions of emergency and non-emergency cases throughout the day are shown in online supplemental eFigure 1. Risks of adverse outcomes were similar between emergency surgeries performed during the night versus day (mortality: ORadj 1.13, 95% CI 0.91 to 1.40, p=0.27; morbidity: ORadj 1.04, 95% CI 0.94 to 1.16, p=0.427).

    Sensitivity analyses

    Propensity score matching and mixed-effects logistic regression analysis in the complete-case cohort confirmed our primary finding (ORadj 1.21, 95% CI 1.06 to 1.39, p=0.005, n=41 414; ORadj 1.34, 95% CI 1.18 to 1.52, n=2 89 480 cases with provider data available). The case delay time was longer in patients who underwent night surgery compared with day surgery (78 (SD 150) vs 17 (SD 50) min, p<0.001). The interaction term between night surgery and case delay in minutes was marginally significant (p=0.055), and the association between night surgery and mortality remained significant in this analysis (ORadj 1.26, 95% CI 1.06 to 1.50, p=0.009, n=1 59 591). Case delay did not mediate the effect of night surgery on mortality. Details of sensitivity analyses are provided in the online supplemental file 1.

    Discussion

    In this large multicentre study of non-cardiac surgical patients, we made the following observations: first, risks of postoperative mortality and morbidity at 30 days were higher among patients who underwent night surgery compared with day surgery. Second, a higher blood transfusion rate and more frequent anaesthesia handovers during night-time partly mediated the effect of night surgery. The adverse effects of night surgery were modified by admission status and acuity level: magnified adverse effects were observed in patients who underwent ambulatory or same-day surgery while the effect was not significant in patients undergoing emergency surgery. These findings suggest that the increased risk of mortality and morbidity after night surgery was not driven by a higher case acuity, but could be attributed to differences in the surgical management of patients operated during night-time.

    Relation to other studies

    Previous studies reported inconclusive results on the effect of operation time on postoperative outcome. More complications after surgery at night were previously described in smaller studies among patients undergoing orthopaedic,17 neurosurgery,18 and plastic surgery,19 laparoscopic cholecystectomy,20 coronary interventions21 and general and vascular surgery.4 22 In the latter study,22 higher mortality was only detected for patients undergoing elective surgery, a finding which we confirm with our study, and was also reported in other studies with negative results in patients who underwent emergency surgery.23 24

    Other studies did not identify harmful effects of night surgery. A single-centre investigation in Germany including non-emergency cases across surgical fields suggested an increased in-hospital mortality only when the surgery was conducted during the early afternoon (13:00 to 17:00 hours).3 In a study that used National Surgical Quality Improvement Program data, night-time surgery in patients undergoing elective surgery was not associated with postoperative morbidity and mortality.4

    Finally, a recent meta-analysis summarised 40 observational studies on the association between night surgery and mortality. The largest study included in this meta-analysis used the National Anesthesia Clinical Outcomes Registry to analyse three patient (age, sex, ASA status) and procedure-related (emergency, surgery type, operation time between 16:00 to 06:59 hours) factors associated with 48-hour mortality after surgery.2 Our study adds the important information that the association between night surgery and increased mortality is robust even when adjusting for important confounders such as case delays and intraoperative risk factors. In addition, we present potentially preventable mediators of the increased risk of mortality and morbidity observed in patients undergoing night surgery: frequent anaesthesia provider handovers and higher rates of intraoperative blood transfusion.

    Meaning of this study

    We present a hypothesis driven study from two major tertiary care hospital networks. In our cohort, patients who underwent night surgery were more often hospitalised and underwent more emergency procedures than those undergoing day surgery. The effect of night surgery on mortality was more pronounced in ambulatory and same-day admitted cases compared with inpatients. This novel observation supports the hypothesis that night surgery itself was associated with adverse outcomes and that its effect was not driven by a higher case acuity or a more severe condition of patients undergoing night surgery. This further suggests that patients who were just admitted to the hospital and who may not have received the appropriate preoperative workup may represent a risk group where it would be clinically meaningful to avoid night surgery. Among emergency patients, we did not find a difference in mortality between day and night cases. The effect of night surgery on adverse outcome may be weaker in emergency patients because other factors such as severity of the condition may be more important, while the emergency surgery has to be performed immediately regardless of the time of day or night.

    The novelty of this study relates to the identification of preventable mediators of the adverse effects of night surgery. Mediation analysis is a tool for providing explanation of an observed association within observational studies.25–27 We identified that the risk of receiving intraoperative blood transfusion was higher among patients undergoing night surgery. The effect was robust when accounting for preoperative anaemia, intraoperative blood loss and an individual provider-related variability in transfusion practice. Mediation analysis revealed that the higher transfusion rate observed during night-time explained 5.9% of the increased mortality and 4.9% of the increased complication risk after night surgery. We speculate that higher transfusion rates independent of patient and procedural risk factors may indicate an early and more liberal treatment of conditions that may occur within the postoperative period, such as a decrease in haemoglobin levels, thereby aiming to prevent the need for further therapy during night-time. Multiple studies have shown that a restrictive rather than liberal transfusion strategy with respect to WHO recommended transfusion thresholds and clinical symptoms of anaemia (transfusion triggers) is associated with improved outcomes after surgery.28–33 Various interventions have been recommended to reduce transfusion rates, such as enhancing a patient’s physiological tolerance of anaemia by optimising oxygenation, decreasing oxygen consumption and ensuring normovolaemia.34 35

    We observed that handovers from one anaesthesia provider to another occurred 2.5 times more often during night surgeries compared with surgeries during day-time. In a recent study published in JAMA, anaesthesia handovers have been identified as a risk factor of increased postoperative mortality and major complications.11 We were able to confirm in our mediation analysis that the higher proportion of handovers contributed to an increased complication risk after night surgery. This may be explained by the fact that during night-time, circulating teams of surgeons, anaesthesia providers and nurses are stretched and less specialised than day teams. It has been shown that liver and thoracic organ transplants do not have adverse outcomes following night surgery which may be due to the typically permanent and highly specialised composition of transplant teams.36–38 Based on our findings, we speculate that low acuity cases such as ambulatory surgeries are particularly prone to errors due to reduced attention and improper communication within changing teams. For the high acuity patients, there may be a higher level of communication and rechecking that prevents failures.

    We also observed significantly longer case delays among patients who underwent night surgery, which could be related to a longer fasting period in patients undergoing surgery at night-time. Fasting prior to surgery leads to a catabolic metabolism that affects a patient’s stress response to surgery and postoperative insulin resistance and increases patient discomfort,39 40 which might add to the negative effects of night surgery on postoperative outcome. Implementation of case-specific fasting guidelines may help decrease the harmful effects of long fasting periods on postoperative outcomes.

    Implications for clinicians and policymakers

    The observed higher proportion of blood transfusions and handovers may reflect differences in the surgical management of patients operated during night-time. Different behaviour patterns of providers during night-time may in some cases rather promote self-interest than patient-centred and individualised care. The observed key factors contributing to a higher risk of night surgery should be modified and adapted to practice patterns during the day to improve outcomes after night surgery.

    In our study, almost 60% of cases performed at night were non-emergency procedures. Of those, 60% were ambulatory or same-day surgeries, which carried the highest risk of mortality attributable to night surgery. In these cases, postponing a surgical procedure should be considered to allow time for both the patient and provider to prepare for surgery. In a recent study, efforts to implement standardised protocols for patient urgency classification and operating room booking aimed at a more selective out-of-hours use of operation rooms for emergency services.41 First results demonstrated higher operating room efficiency during standard hours (increased case time during the day), while operation time at night decreased by 26%.41 Adapted to the local conditions, further hospitals should evaluate the implementation of standardised scheduling tools for non-emergency patients. In addition, if night surgery has to be conducted, we suggest that blood transfusion protocols should be rigorously implemented, and handovers be minimised.

    There are several limitations to our study. Patients who underwent night surgery were more often inpatients and emergency cases and may be generally sicker than patients undergoing surgery during standard hours. To address confounding related to these differences, we used an interaction analysis which demonstrated a greater effect of night surgery on mortality among patients undergoing ambulatory/same-day surgery compared with inpatients, while no association between night surgery and mortality was found in a subgroup of emergency patients. We used several sensitivity analyses such as propensity score matching and repeating the primary analysis in the complete-case cohort, and the similar results were confirmatory. In addition, path mediation models have been described as a more complex form of multiple regression models, which still cannot establish causality of an effect (as compared with experimental studies) but can be used to determine whether hypotheses from observational data are plausible.25 42 However, potential unmeasured confounders may have affected results of this observational study.

    Conclusion

    We demonstrate that surgery at night was associated with an increased risk of mortality and morbidity in a large multicentre cohort. Patients who underwent ambulatory or same-day surgery were particularly vulnerable to the adverse effect of night surgery. Higher exposures to blood transfusions and anaesthesia handovers during night-time partly explained the increased complication risk. Based on our findings, the risk of night surgery did not appear to be driven by a higher case acuity but could partly be attributed to a different surgical management of patients operated during night-time.

    Data availability statement

    Data are available on reasonable request. Due to the sensitive nature of the data collected for this study, requests to access the dataset from qualified researchers trained in human subjects research and confidentiality may be sent to Matthias Eikermann at meikerma@bidmc.harvard.edu.

    Ethics statements

    References

    1. ↵
      2. Cortegiani A ,
      3. Gregoretti C ,
      4. Neto AS , et al
      . Association between night-time surgery and occurrence of intraoperative adverse events and postoperative pulmonary complications. Br J Anaesth 2019;122:361–9.doi:10.1016/j.bja.2018.10.063 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30770054
      OpenUrlPubMed
    2. ↵
      2. Whitlock EL ,
      3. Feiner JR ,
      4. Chen L-L
      . Perioperative mortality, 2010 to 2014: a retrospective cohort study using the National anesthesia clinical outcomes registry. Anesthesiology 2015;123:1312–21.doi:10.1097/ALN.0000000000000882 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26492481
      OpenUrlCrossRefPubMed
    3. ↵
      2. Kork F ,
      3. Spies C ,
      4. Conrad T , et al
      . Associations of postoperative mortality with the time of day, week and year. Anaesthesia 2018;73:711–8.doi:10.1111/anae.14228 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29473682
      OpenUrlPubMed
    4. ↵
      2. Kelz RR ,
      3. Freeman KM ,
      4. Hosokawa PW , et al
      . Time of day is associated with postoperative morbidity: an analysis of the National surgical quality improvement program data. Ann Surg 2008;247:544–52.doi:10.1097/SLA.0b013e31815d7434 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18376202
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      2. Anderson C ,
      3. Sullivan JP ,
      4. Flynn-Evans EE , et al
      . Deterioration of neurobehavioral performance in resident physicians during repeated exposure to extended duration work shifts. Sleep 2012;35:1137–46.doi:10.5665/sleep.2004 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22851809
      OpenUrlPubMedWeb of Science
    6. ↵
      2. Barger LK ,
      3. Ayas NT ,
      4. Cade BE , et al
      . Impact of extended-duration shifts on medical errors, adverse events, and attentional failures. PLoS Med 2006;3:e487. doi:10.1371/journal.pmed.0030487 pmid:http://www.ncbi.nlm.nih.gov/pubmed/17194188
      OpenUrlCrossRefPubMed
    7. ↵
      2. Landrigan CP ,
      3. Rothschild JM ,
      4. Cronin JW , et al
      . Effect of reducing interns' work hours on serious medical errors in intensive care units. N Engl J Med 2004;351:1838–48.doi:10.1056/NEJMoa041406 pmid:http://www.ncbi.nlm.nih.gov/pubmed/15509817
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Gold DR ,
      3. Rogacz S ,
      4. Bock N , et al
      . Rotating shift work, sleep, and accidents related to sleepiness in hospital nurses. Am J Public Health 1992;82:1011–4.doi:10.2105/ajph.82.7.1011 pmid:http://www.ncbi.nlm.nih.gov/pubmed/1609900
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      2. Howard SK ,
      3. Rosekind MR ,
      4. Katz JD , et al
      . Fatigue in anesthesia: implications and strategies for patient and provider safety. Anesthesiology 2002;97:1281–94.doi:10.1097/00000542-200211000-00035 pmid:http://www.ncbi.nlm.nih.gov/pubmed/12411816
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      2. von Elm E ,
      3. Altman DG ,
      4. Egger M , et al
      . The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet 2007;370:1453–7.doi:10.1016/S0140-6736(07)61602-X pmid:http://www.ncbi.nlm.nih.gov/pubmed/18064739
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      2. Jones PM ,
      3. Cherry RA ,
      4. Allen BN , et al
      . Association between handover of anesthesia care and adverse postoperative outcomes among patients undergoing major surgery. JAMA 2018;319:143–53.doi:10.1001/jama.2017.20040 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29318277
      OpenUrlPubMed
    12. ↵
      2. Chan DXH ,
      3. Sim YE ,
      4. Chan YH , et al
      . Development of the combined assessment of risk encountered in surgery (cares) surgical risk calculator for prediction of postsurgical mortality and need for intensive care unit admission risk: a single-center retrospective study. BMJ Open 2018;8:e019427. doi:10.1136/bmjopen-2017-019427 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29574442
      OpenUrlAbstract/FREE Full Text
    13. ↵
      2. Buis ML
      . Direct and indirect effects in a logit model. Stata J 2010;10:11–29.doi:10.1177/1536867X1001000104 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22468140
      OpenUrlPubMedWeb of Science
    14. ↵
      2. Grotta A ,
      3. Bellocco R
      . A review of mediation analysis in Stata: principles, methods and applications 2013. Available: https://www.stata.com/meeting/italy13/abstracts/materials/it13_grotta.pdf
    15. ↵
      2. Hadland SE ,
      3. Rivera-Aguirre A ,
      4. Marshall BDL , et al
      . Association of pharmaceutical industry marketing of opioid products with mortality from Opioid-Related overdoses. JAMA Netw Open 2019;2:e186007–e07.doi:10.1001/jamanetworkopen.2018.6007 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30657529
      OpenUrlPubMed
    16. ↵
      2. Ng PY ,
      3. Ng AK-Y ,
      4. Subramaniam B , et al
      . Association of preoperatively diagnosed patent foramen ovale with perioperative ischemic stroke. JAMA 2018;319:452–62.doi:10.1001/jama.2017.21899 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29411032
      OpenUrlCrossRefPubMed
    17. ↵
      2. Ricci WM ,
      3. Gallagher B ,
      4. Brandt A , et al
      . Is after-hours orthopaedic surgery associated with adverse outcomes? A prospective comparative study. J Bone Joint Surg Am 2009;91:2067–72.doi:10.2106/JBJS.H.00661 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19723981
      OpenUrlAbstract/FREE Full Text
    18. ↵
      2. Linzey JR ,
      3. Burke JF ,
      4. Sabbagh MA , et al
      . The effect of surgical start time on complications associated with neurological surgeries. Neurosurgery 2018;83:501–7.doi:10.1093/neuros/nyx485 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29040696
      OpenUrlPubMed
    19. ↵
      2. Lee K-T ,
      3. Mun G-H
      . Is after-hours free-flap surgery associated with adverse outcomes? J Plast Reconstr Aesthet Surg 2013;66:460–6.doi:10.1016/j.bjps.2012.12.007 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23395287
      OpenUrlPubMed
    20. ↵
      2. Phatak UR ,
      3. Chan WM ,
      4. Lew DF , et al
      . Is nighttime the right time? risk of complications after laparoscopic cholecystectomy at night. J Am Coll Surg 2014;219:718–24.doi:10.1016/j.jamcollsurg.2014.05.009 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25172046
      OpenUrlPubMed
    21. ↵
      2. Glaser R ,
      3. Naidu SS ,
      4. Selzer F , et al
      . Factors associated with poorer prognosis for patients undergoing primary percutaneous coronary intervention during off-hours: biology or systems failure? JACC Cardiovasc Interv 2008;1:681–8.doi:10.1016/j.jcin.2008.08.020 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19463384
      OpenUrlAbstract/FREE Full Text
    22. ↵
      2. Kelz RR ,
      3. Tran TT ,
      4. Hosokawa P , et al
      . Time-of-day effects on surgical outcomes in the private sector: a retrospective cohort study. J Am Coll Surg 2009;209:434–45.doi:10.1016/j.jamcollsurg.2009.05.022 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19801316
      OpenUrlPubMed
    23. ↵
      2. van Zaane B ,
      3. van Klei WA ,
      4. Buhre WF , et al
      . Nonelective surgery at night and in-hospital mortality: prospective observational data from the European surgical outcomes study. Eur J Anaesthesiol 2015;32:477–85.doi:10.1097/EJA.0000000000000256 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26001104
      OpenUrlCrossRefPubMed
    24. ↵
      2. Zafar SN ,
      3. Libuit L ,
      4. Hashmi ZG , et al
      . The sleepy surgeon: does night-time surgery for trauma affect mortality outcomes? Am J Surg 2015;209:633–9.doi:10.1016/j.amjsurg.2014.12.015 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25681253
      OpenUrlPubMed
    25. ↵
      2. Rudolph KE ,
      3. Goin DE ,
      4. Paksarian D , et al
      . Causal mediation analysis with observational data: considerations and illustration examining mechanisms linking neighborhood poverty to adolescent substance use. Am J Epidemiol 2019;188:598–608.doi:10.1093/aje/kwy248 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30561500
      OpenUrlPubMed
    26. ↵
      2. Camacho-Soto A ,
      3. Warden MN ,
      4. Searles Nielsen S , et al
      . Traumatic brain injury in the prodromal period of Parkinson's disease: a large epidemiological study using Medicare data. Ann Neurol 2017;82:744–54.doi:10.1002/ana.25074 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29024046
      OpenUrlPubMed
    27. ↵
      2. Atlantis E ,
      3. Shi Z ,
      4. Penninx BJWH , et al
      . Chronic medical conditions mediate the association between depression and cardiovascular disease mortality. Soc Psychiatry Psychiatr Epidemiol 2012;47:615–25.doi:10.1007/s00127-011-0365-9 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21384119
      OpenUrlPubMed
    28. ↵
      2. Hébert PC ,
      3. Wells G ,
      4. Blajchman MA , et al
      . A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. transfusion requirements in critical care Investigators, Canadian critical care Trials Group. N Engl J Med 1999;340:409–17.doi:10.1056/NEJM199902113400601 pmid:http://www.ncbi.nlm.nih.gov/pubmed/9971864
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      2. Hajjar LA ,
      3. Vincent J-L ,
      4. Galas FRBG , et al
      . Transfusion requirements after cardiac surgery: the TRACS randomized controlled trial. JAMA 2010;304:1559–67.doi:10.1001/jama.2010.1446 pmid:http://www.ncbi.nlm.nih.gov/pubmed/20940381
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      2. Carson JL ,
      3. Terrin ML ,
      4. Noveck H , et al
      . Liberal or restrictive transfusion in high-risk patients after hip surgery. N Engl J Med 2011;365:2453–62.doi:10.1056/NEJMoa1012452 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22168590
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      2. Althoff FC ,
      3. Neb H ,
      4. Herrmann E , et al
      . Multimodal patient blood management program based on a Three-pillar strategy: a systematic review and meta-analysis. Ann Surg 2019;269:794–804.doi:10.1097/SLA.0000000000003095 pmid:http://www.ncbi.nlm.nih.gov/pubmed/30418206
      OpenUrlPubMed
    32. ↵
      1. WH A
      . Availability, safety and quality of blood products, 2010. Available: http://apps.who.int/gb/ebwha/pdf_files/WHA63/A63_R12-en.pdf
    33. ↵
      2. Farmer SL ,
      3. Towler SC ,
      4. Leahy MF , et al
      . Drivers for change: Western Australia patient blood management program (WA PBMP), world health assembly (WHA) and Advisory Committee on blood safety and availability (ACBSA). Best Pract Res Clin Anaesthesiol 2013;27:43–58.doi:10.1016/j.bpa.2012.12.007 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23590915
      OpenUrlPubMed
    34. ↵
      2. Meybohm P ,
      3. Richards T ,
      4. Isbister J , et al
      . Patient blood management bundles to facilitate implementation. Transfus Med Rev 2017;31:62–71.doi:10.1016/j.tmrv.2016.05.012 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27317382
      OpenUrlPubMed
    35. ↵
      2. Leahy MF ,
      3. Roberts H ,
      4. Mukhtar SA , et al
      . A pragmatic approach to embedding patient blood management in a tertiary hospital. Transfusion 2014;54:1133–45.doi:10.1111/trf.12362 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23927725
      OpenUrlPubMed
    36. ↵
      2. Thuluvath PJ ,
      3. Amjad W ,
      4. Savva Y , et al
      . Survival outcomes are not affected when liver transplant surgery is done at night, during weekends, or summer months. Transplant Direct 2019;5:e449. doi:10.1097/TXD.0000000000000887 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31165084
      OpenUrlPubMed
    37. ↵
      2. Halliday N ,
      3. Martin K ,
      4. Collett D , et al
      . Is liver transplantation ‘out-of-hours’ non-inferior to ‘in-hours’ transplantation? A retrospective analysis of the UK Transplant Registry. BMJ Open 2019;9:e024917. doi:10.1136/bmjopen-2018-024917
    38. ↵
      2. George TJ ,
      3. Arnaoutakis GJ ,
      4. Merlo CA , et al
      . Association of operative time of day with outcomes after thoracic organ transplant. JAMA 2011;305:2193–9.doi:10.1001/jama.2011.726 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21632483
      OpenUrlCrossRefPubMed
    39. ↵
      2. Gustafsson UO ,
      3. Scott MJ ,
      4. Schwenk W , et al
      . Guidelines for perioperative care in elective colonic surgery: Enhanced Recovery After Surgery (ERAS(®)) Society recommendations. World J Surg 2013;37:259–84.doi:10.1007/s00268-012-1772-0 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23052794
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      2. Longchamp A ,
      3. Harputlugil E ,
      4. Corpataux J-M , et al
      . Is overnight fasting before surgery too much or not enough? how basic aging research can guide preoperative nutritional recommendations to improve surgical outcomes: a mini-review. Gerontology 2017;63:228–37.doi:10.1159/000453109 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28052287
      OpenUrlPubMed
    41. ↵
      2. Sandbaek BE ,
      3. Helgheim BI ,
      4. Larsen OI , et al
      . Impact of changed management policies on operating room efficiency. BMC Health Serv Res 2014;14:224–24.doi:10.1186/1472-6963-14-224 pmid:http://www.ncbi.nlm.nih.gov/pubmed/24885869
      OpenUrlPubMed
    42. ↵
      2. Streiner DL
      . Finding our way: an introduction to path analysis. Can J Psychiatry 2005;50:115–22.doi:10.1177/070674370505000207 pmid:http://www.ncbi.nlm.nih.gov/pubmed/15807228
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • FCA, LJW and PR contributed equally.

    • Contributors Study concept and design: FCA, PR, PS, SDG, MSS, XX, HC, OA, GAB, TTH, ME. Acquisition, analysis or interpretation of data: All authors. Drafting of the manuscript: FCA, PR, LJW, ME. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: XX, TTH, FCA, PR, LJW, PS, GAB, MSS, ME. Obtained funding: ME. Administrative, technical or material support: All authors. Study supervision: XX, TTH, GAB, ME. All authors had full access to the data in the study and can take responsibility for the integrity of the data and the accuracy of analyses.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests ME has received unrestricted funds from philanthropic donors Jeffrey and Judith Buzen during the study period and has received grants for investigator-initiated trials not related to this manuscript from Merck & Co and serves as a consultant on the advisory board of Merck & Co. ME holds equity of Calabash Bioscience Inc. and is an Associate Editor of the British Journal of Anaesthesia.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.