- Email: [email protected]
Perioperative mortality related to anesthesia within 48 h and up to 30 days following surgery: A retrospective cohort study of 11,562 anesthetic procedures
Journal of Clinical Anesthesia 49 (2018) 79–86
Contents lists available at ScienceDirect
Journal of Clinical Anesthesia journal homepage: www.elsevier.com/locate/jclinane
Original Contribution
Perioperative mortality related to anesthesia within 48 h and up to 30 days following surgery: A retrospective cohort study of 11,562 anesthetic procedures
T
⁎
Luciana C. Stefani, MD, PhDa,b, , Patricia W. Gamermann, MDb, Amanda Backof, MDb, Fernanda Guollo, MDb, Rafael M.J. Borgesc, Adriana Martin, MDd, Wolnei Caumo, MD, PhDa,b,e, Elaine A. Felix, MD, PhDa,b a
Department of Surgery, School of Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil Serviço de Anestesia e Medicina Perioperatória, Hospital de Clínicas de Porto Alegre (HCPA), Porto Alegre, Brazil School of Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil d Cleveland Clinic, Cleveland, USA e Pain and Neuromodulation Laboratory, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Perioperative mortality Anesthesia-related death Surgical risk In-hospital mortality
Study objective: Studying postoperative in-hospital mortality is crucial to the understanding of the perioperative process failures and to the implementation of strategies to improve patient outcomes. We intend to classify the causes of perioperative deaths up to 30 days after procedures requiring anesthesia and to evaluate the risk factors for early (48 h) or late (30 day) mortality. Design: Retrospective cohort study. Setting: A quaternary University Hospital from South Brazil. Patients: The information related to the perioperative care was collected from surgeries performed between January 2012 and December 2011. Interventions: None (observational study). Measurements: Three anesthesiologists classified the causes of deaths according to the ANZCA (Australian and New Zealand College of Anesthetists) classification, used in the report of Anesthesia-Related Mortality in Australia since 1985, which defines eight death categories. The risk factors for early or late death were analyzed in a regression model. Main results: 11.562 surgeries were performed, with a mortality incidence of 2.75% within 30 days (319 deaths). Most deaths were inevitable (50.7%), as they were related to advanced illnesses and would occur regardless of anesthetic or surgical procedures. The second most common cause was related to surgical complications (25%). The death rate having anesthesia as a likely contributor was 1.72:10.000 procedures, and as a potential contributor 7.78:10.000. These deaths occurred significantly earlier (< 48 h) when compared to deaths from other causes. Transoperative vasopressor, extremes of age and out-of-hour surgery were independent variables associated to early deaths. Conclusions: The study confirms that postoperative mortality in which anesthesia was involved occurred earlier in the perioperative period. In addition, it was revealed that this involvement of anesthesia as a morbidity contributor shows higher frequency when considering the anesthesiologist perioperative role, and when assessing the mortality in the long term (30 days).
1. Introduction Global perioperative risk is multifactorial. It depends on the interaction between anesthesia, patient's clinical conditions, and specific aspects of the surgery [1]. Preoperative stratification comprehends a ⁎
deep knowledge about the patient, the identification of precise risks, and to the adequate trans- and postoperative management. Postoperative morbidity and mortality can be assessed in different moments: in the transoperative period, in the immediate postoperative period, in 48 h, 30 days, or in one year or more following surgery.
Corresponding author at: Rua Ramiro Barcelos, 2350, 90035-003 Porto Alegre, RS, Brazil. E-mail addresses: [email protected] (L.C. Stefani), [email protected] (P.W. Gamermann), [email protected] (W. Caumo), [email protected] (E.A. Felix).
https://doi.org/10.1016/j.jclinane.2018.06.025 Received 20 February 2018; Received in revised form 2 June 2018; Accepted 8 June 2018 0952-8180/ © 2018 Elsevier Inc. All rights reserved.
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Fig. 1. Flowchart of the methodology used in the study.
Committee of Grupo de Pesquisa e Pós-Graduação from Hospital de Clínicas de Porto Alegre – Brazil (Chairperson Prof. Patricia Prolla) on 22 may 2014 (protocol number 14-0252.). The Hospital de Clínicas de Porto Alegre is a reference university quaternary care institution in southern Brazil. It holds 842 beds and performs around 20,000 surgeries a year, serving all age groups. The Anesthesiology and Perioperative Medicine Service comprises a permanent team of 83 anesthesiologists and 40 other resident doctors certified by the Brazilian Anesthesiology Society. The institution's information management system stores data from all surgical procedures and patients who underwent them. Using this system, we collected data from the deaths that occurred between anesthesia and up to 30 days post-surgery, from January 1st, 2012, to December 31st, 2013. The outcome was in-hospital death up to 30 days. When the patients overstay the 30-day period, the following days were not included in the analysis. Data were obtained by reviewing electronic and paper records (anesthesia files). The researchers were resident doctors trained to find information related to preoperative comorbidity, surgery and anesthesia, as well as postoperative complications, and details on the event of death. After data collection, three of the anesthesiologists
Statistics vary according to the time of assessment, but it is known that, usually, a small group of high-risk patients accounts for a greater part of the deaths and significantly longer hospitalizations [2, 3]. Therefore, identifying and signaling high-risk patients is essential to provide appropriate care, to implement strategies for preventing complications, and to precisely allocate resources. With the advancement of anesthesia and surgery techniques over the last 50 years, perioperative morbidity and mortality rates decreased due to improvements in safety, enhancements in monitoring techniques, the development and adoption of care protocols, and systematic measures for risk reduction [4–6]. Considering the evolution of anesthesia safety and its role in the perioperative medicine, it's reasonable to analyze the impact of anesthesia act in the postoperative mortality during a larger period of observation. This study aims to assess the causes of perioperative deaths within 30 days following procedures requiring anesthesia. We also intend to compare the risk factors associated to early (48 h) or late deaths (30 days). 2. Methods This retrospective cohort study had ethical approval by the Ethical 80
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
authors of this study (LS, AB, FG) met to discuss and classify the deaths and their probable causes. The ANZCA (Australian and New Zealand College of Anesthetists) [7] classification served as a background source in this process. This report of anesthesia-related mortality in Australia has been done since 1985. Its methodology defines eight death categories – the first three categories are considered deaths attributable to anesthesia, in the other three anesthesia played no part and deaths are related to surgical factors or other factors beyond the control of the physicians, and the last two categories are classified as unassessable deaths. For the deaths related somehow to anesthesia, there are 24 subcategories. The complete tool is detailed in Supplements 1 and 2. The strobe guideline for observational studies was followed and could be accessed in supplement.
Table 1 Characteristics of patients and procedures that resulted in death up to and after 48 h following surgery. Death < 48 h
2.1. Statistical analysis The results are presented as interquartile intervals, or absolute frequencies. Variables related to early and late deaths were compared using the chi-square test, Fisher's exact test, or Student's t-test for continuous data. A logistic regression model was used to identify which factors were independently associated with early deaths. Clinically significant variables or variables with p < 0.1 in the univariate analysis were included in the model. The significance level was set at p < 0.05 (bicaudal). Data analysis was performed using the Statistical Package for the Social Sciences software (SPPS), version 20.0 for MAC (SPSS Inc., Chicago, IL).
Initially, a total of 40.505 procedures were identified for the analyzed period. Their duplicates were discarded, as well as diagnostic and procedures under local anesthesia. The valid procedures were then arranged together and categorized into minor, intermediate, or major. These categories were based on previous study [8] and discussed among the institution's surgical specialists. The analysis took into consideration only one procedure per patient, choosing the largest one. Fig. 1 shows the study flowchart. A total of 11.562 surgeries were analyzed. The incidence of postoperative mortality within 48 h was 0.65%, while the incidence of postoperative mortality in patients hospitalized for up to 30 days was 2.75%. Most of these deaths were concentrated in patients ≥ ASA III (86.9%) submitted to non-elective surgeries, either urgent or emergent (55%). Table 1 presents patient characteristics, preoperative risk conditions, and information about the nature of the surgery, as well as surgery dimension, and the anesthetic technique used. It compares early (< 48 h) and late (> 48 h) deaths. The procedure with the highest rate of deaths was exploratory laparotomy (22.1% of total deaths), followed by colorectal resections (5.9%), and upper and lower limb amputations (5.6%). The exploratory laparotomy rate was higher than other procedures, 71 deaths in 461 procedures (15.4%). A logistic regression model identified the presence of extremes of age [OR 1,96 IC 1,04-3,69], transoperative vasopressors use [OR 3,36 IC 1,73–6,53) and out-of-hour surgery [OR 3,73 IC 1,36–10,18) as factors independently associated with early mortality in comparison with late mortality (Table 2). The most frequent causa mortis both in early and late deaths was sepsis (p = 0.06), followed by bleeding in early deaths (p < 0.01), and advanced neoplasm in early and late deaths (p = 0.42) (Fig. 2). A consensual qualitative analysis of the sample's death cases identified the causes of deaths and their relation with the anesthesia, the surgical procedure per se, the patient conditions or the combining of these factors. It has been shown that early deaths are more probably related to anesthesia than deaths over 48 h (p = 0.02) (Table 3). Most
p
n = 76
%
n = 243
%
Age (years) 0–1 2–15 16–35 36–50 51–64 65–79 ≥80
7 1 2 3 17 35 11
9.3 1.3 2.7 4.0 21.3 46.7 14.7
16 8 14 24 67 84 30
6.5 3.3 5.7 9.8 28.2 34.3 12.2
0.42 0.37 0.27 0.11 0.23 0.06 0.62
ASA I II III IV V
0 5 22 37 12
0 6.7 29 49.3 16
2 35 120 75 11
0.8 14.3 49.4 31 4.5
0.42 0.09 < 0.01 < 0.01 < 0.01
Degree of surgerya Minor Intermediary Major
3 17 56
4 22 73.3
33 77 133
13.5 31.4 55.1
0.02 0.13 0.01
Nature Urgent Elective
50 25
66.7 33.3
124 121
50.6 49
0.02
13
17.1
11
4.5
0.01
63
83
232
95.5
Perioperative risk situation Ischemic heart disease Renal insufficiency [cr > 2] Functional capacity < 4METs Cardiac insufficiency DM, insulin-dependent Hemodynamic instabilitya Bowel obstruction Preoperative anemia Preoperative sepsis Vasoactive drug Transfusion
8 10 65 10 6 30 12 20 17 50 28
10.7 13.3 86.7 13.3 8 40 16 26.7 22.7 66 37.3
27 33 190 29 29 43 35 83 33 68 31
11 13.5 77.6 11.8 11.8 17.6 14.3 33.9 13.5 27.8 12.7
0.92 1.0 0.09 0.76 0.37 < 0.01 0.69 0.23 0.06 < 0.01 < 0.01
Anesthesia type General anesthesia Regional anesthesia Sedation Combined anesthesia
72 2 0 0
97.3 2.7 0 0
199 27 3 16
81.2 11 1.2 6.5
< 0.01 0.03 0.32 0.02
Schedule Out-of-hour surgery [after midnight] Surgery before midnight
3. Results
Death > 48 h
a
Uncontrolled hypotension, or hypoperfusion signs.
deaths were considered inevitable (50.9%), meaning that they would occur irrespective of the anesthetic or surgical act, but due to the patient's unfavorable conditions. The second category identified were deaths related to surgery in which anesthesia was not a contributor (26,1%). Two cases were classified as related to anesthesia or other factors controlled by the anesthetist (category 2: Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist). These two deaths were associated with cardiac arrest at anesthetic induction in the first case and with immediate postoperative hypoxemia and cardiac arrest in the second case. Nine cases were categorized as having been caused by both surgical or anesthetic acts. Among these deaths partially related to anesthesia (category 3), the main contributing factors were hemorrhagic shock (38.5%), followed by sepsis (23%), thromboembolic phenomena - such as acute myocardial infarction (15.38%), and complications associated to hypoxemia - such as aspiration and pneumonia (15.38%). The description of all cases possible related to anesthesia is in Table 4.
81
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Table 2 Logistic regression of risk factors for deaths up to 48 h vs. after 48 h. C coeficiente (B)
Standard error
p value
Odds ratio
95% IC
Variables related to patients Extreme age ASA > 3 Low functional capacityb Pre or transoperative transfusion Transoperative continuous vasoactive drugs Hemodynamic instability after surgeryc
0,67 0,64 −0,45 0,70 1,21 0,48
0,32 0,51 0,41 0,38 0,33 0,34
0,03 0,35 0,27 0,067 0,00 0,15
1,96 1,90 0,63 2,02 3,36 1,62
1,04–3,69 0,64–5,6 0,28–1,42 0,95–4,31 1,73–6,53 0,82–3,18
Variables related to surgery Surgery duration > 150 min Emergency operation Out-of-hour surgerya Anesthesia type
0,28 0,24 1,31 1,47
0,33 0,34 0,51 0,77
0,40 0,46 0,01 0,058
1,32 1,27 3,73 4,35
0,68–2,56 0,65–2,4 1,36–10,18 0,95–19,91
Odds ratios were calculated with the use of multivariable logistic regression and they are for the group of patients that died before 48 h as compared with those that died after 48 h from the surgery. An odds ratio of > 1 represents a greater likelihood of death in < 48 h after surgery. Listed are variables with p < 0,1 in the multivariable model. Risk factors: age (0–1 or +65 years vs. 2–64 years), ASA PS (3, 4, 5 vs. 1, 2), low functional capacity/transfusion, vasoactive drugs, hemodynamic instability (yes vs. no), Anesthesia type (general vs. regional or combined or sedation). a Unsafe working hours (after midnight to seven). b Capacity < 4 METs [metabolic milliequivalent]. c Intensive care resource recovery or hemodynamic monitoring after surgery.
Our study has two main contributions: first, we looked deeper into the anesthesia influence on postoperative deaths up to 30 days reviewing all cases in a consensual manner and classifying their probable causes. Among the identifiable causes, the patients' conditions were the most relevant contributing factor, resulting in inevitable deaths that would occur regardless of surgery or anesthesia (50%). The second most common cause identified was related to complications from the surgical procedure (25%). Although, anesthesia was almost certainly involved in 1.72:10.000 procedures, and was a potential contributor in 7.78:10.000 cases. This is significantly higher than the most recent data, published from an Australian audit, which was around 0.17:10.000 [7]. However, this is a national audit from five Australian states, and as it is stated on the conclusion, they do not have available methods to verify whether all anesthesia and/or sedation deaths are reported under that provision. Furthermore, they evaluate only deaths up to 24 h after surgery, and we evaluate deaths up to 30 days after surgery. Otherwise, a recently audit from deaths observed within the 30-day period following anesthesia declared to the SHAM insurance (Société Hospitalière d'Assurance Mutuelle-Paris) over the last 10 years found concordant levels to our cohort. They showed an incidence of 0.88 per 10.000 deaths related to anesthesia in ambulatory patients [15].
4. Discussion This study identified a perioperative mortality rate of 0.64% within 48 h after the procedure and 2.74% in patients hospitalized for up to 30 days following surgery. In general, most deaths occurred later after surgery, in critical patients (ASA ≥ III) submitted to urgent surgeries and were associated with complications such as sepsis and advanced neoplastic disease. Our results are consistent with postoperative death occurrences in hospitalized patients in developed countries, as shown by Fecho et al. in 2008 [9] who has presented a similar death rate: 0.57% within 48 h and 2.1% up to 30 days. Ghaferi et al. [10], however, assessed mortality rates in different American hospitals in 2009 and found values that ranged from 3.5% to 6.9%. This is consistent with a recent European study [11] that compared various countries for perioperative mortality of inpatients, pointing out the negative correlation between mortality and the Human Development Index of each country. Despite recent advances in quality improvements and safety of the surgical patient [12, 13] there is scarce recent research focusing on anesthesia safety and on the role of the anesthetist as a perioperative physician capable to optimize outcomes. Therefore, postoperative mortality in 48 h and 30 days can be used to evaluate the anesthesia performance beyond the transoperative period [14].
Fig. 2. Causa mortis associated with deaths < 48 h or > 48 h. Bars followed by (*) present significant differences.
82
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Table 3 Incidence of deaths and categories analyzed in the perioperative period of 11,562 procedures. Death category according to the ANZCA classification
Death attributable to anesthesia I) Where it is reasonably certain that death was caused by the anesthesia or other factors under the control of the anesthetist II) Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist. III) Where it is reasonably certain that death was caused by both surgical and anesthesia factors. Death not attributable to anesthesia IV) Death where the administration of the anesthesia is not contributory and surgical or other factors are implicated. V) Inevitable death, which would have occurred irrespective of anesthesia or surgical procedures. VI) Incidental death which could not reasonably be expected to have been foreseen by those looking after the patient, was not related to the indication for surgery and was not due to factors under the control of the anesthetist or surgeon. VII) Those that cannot be assessed despite considerable data but where the information is conflicting or key data are missing. VIII) Cases that cannot be assessed because of inadequate data.
Death < 48 h n = 76
Death > 48 h n = 243
Total (%)
p
– 1
– 1
– 2 (0.6%)
– 0.37
5
4
8 (2,5%)
0.02
25 41 1
58 122 51
0.13 0.55 0.00
2 1
6 1
83 (26,1%) 162 (50,9%) 52 (16.4%) 8 (2.5%) 2 (0.6%)
0.92 0.37
adequate management of the surgical schedule, focusing in avoiding out-of-hour's procedures that could have been done in diurnal shift. The anesthetists could serve as a median point between nurses and surgeons in this patient priority decision-making process [11, 19]. The major strength of our study is the exhaustive revision of the pre, trans e postoperative care period in order to identify possible death risk factors linked somehow to the anesthesia act. Preoperatively suboptimal care related to inadequate patient evaluation or suboptimal management of postoperative complications occurred in 63% of our deaths related to anesthesia. Thus, to improve the anesthesia safety, institutions should prioritized the perioperative assistance as much as the intraoperative period [6, 20]. The information related to surgical procedures and the general patients' data were collected from the hospital's information management system and the researchers individually assessed the medical files. The long period of assessment (30 days) allowed us to evaluate late complications and identify their possible associated causes. Using a tool to classify deaths into categories based on a consensus among three anesthetists reduces the subjectivity of the classification, which could be a problem in studies based on a single reviewer. Nevertheless, it is not possible to evaluate the direct influence of teamwork, failure in communication, or leadership problems during the intercurrences in a retrospective study. These non-technical skills are always present on the root cause analysis of any adverse event and should not be minimized. Methodological limitations include the observational design, which is exploratory and dependent on the quality of the data registry. This means that transoperative data rely on the commitment of each anesthetist to the registry process. Additionally, the data is originated from a single institution and may not reflect other scenarios.
We could point out the lack of deaths caused reasonably certainly by anesthesia in our cohort. Although, in two cases, the anesthesia was involved in negative outcomes once the postoperative period handling and the incorrect judgment of patients' particular clinical conditions led to instabilities during the induction and extubation phases. We also found that other deaths somehow related to, but not caused by anesthesia, had influence on both surgical and anesthetic factors. Probably, these cases wouldn't have been even recognized if other less rigid classification tool was adopted, since bleeding, shock, sepsis, and severe ileus were likely considered common surgical complications without anesthesia's involvement. The detailed cases revision revealed that the perioperative commitment of the anesthesiologist seems suboptimal, since the lack of preoperative compensation or the development of postoperative complications could have been avoided, in some of the cases, by reducing the perioperative care fragmentation. The second contribution of our study is the confirmation that the vast majority of deaths with any connection to the anesthetist's performance were of early occurrence (p < 0.01). We found that 7:76 deaths (9,3%) in < 48 h were related to anesthesia; compared to 5:246 from late deaths (2,4%, p ≤ 0,01). The early deaths often reflect the comorbidities and physiological disorders of the patient, the quality and complexity of surgical care, or the risks of anesthesia. However, it cannot be used to compare one site, facility, or country with another without appropriate, validated, and time-consuming risk adjustment [16]. Independently of the cause, the risk factors significantly associated to early deaths were extremes of age (OR 1,96 IC 1,04–3,69), transoperative continuous vasoactive drugs (OR 3,36 IC 1,73–6,53) and out-of-hour surgeries (OR 3,73 IC 1,36–10,28). Extremes of age are associated to physiological fragility and have been identified as a risk factor for perioperative adverse events [10,18,19]. In our study, infants (0–1 year old) and older people (65+ years old) had higher rates of early postoperative mortality than people ranging 2–64 years old. The procedures under these populations should have an extra attention from the caregivers and be prioritized in the list of emergent surgeries. Moreover, transoperative vasoactive drug use reflects the patient severity or the complexity of surgical procedure and it could be associated to any kind of shock, therefore it could be considered an epiphenomenon. Also, performing a surgery during the night (from19:00 h to 07:00 h) had clear impact on outcomes: prolonged hours of work and night shifts bring with them the well documented risks to the performance of the surgical and anesthetic team, and therefore, the quality of patient care provided. The potentially serious consequence of fatigue increases the risk of serious errors that can lead to death or serious morbidity. The operational efficiency of a surgical center involves the
5. Conclusions This study showed that the majority of late postoperative deaths were inevitable (50.7%), as they were related to advanced illnesses and would happen regardless of the anesthetic or surgical acts. Differently from traditional studies on deaths related to anesthesia, we looked deeper and longer into the perioperative path, and found that anesthesia is almost certainly involved in 1.72:10.000 procedures, and was a potential contributor in 10.37:10.000 cases. Furthermore, death cases in which the anesthetic contribution was recognized were significantly premature (< 48 h) when compared to other investigated causes. This study reinforced the need to implement specific multiprofessional teams dedicated to the pre, trans and postoperative care of the high-risk patients, focusing on risk management and reducing the failure to rescue during the perioperative journey.
83
ASA
Procedure
Time from operation to death
Consensus subcategory
Possible prevention
84
3
3
4
75
62
4
56
73
Cholecystectomy
Bypass femoropopliteus
Bypass femoropopliteus
Bypass with stent
24 h after surgery
24 h after surgery
48 h
< 24 h after surgery
Failure to comply with minimum monitoring standards or lack of advanced monitoring when indicated
Death resulted from inappropriate management or lack of active intervention by the anesthetist in the postoperative period. Patient clinical conditions were significant for the occurrence of death related to anesthesia Unsuitable handling of unexpected occurrences during anesthesia
Unsuitable handling of unexpected occurrences during anesthesia. Patient clinical conditions were significant for the occurrence of death related to anesthesia
Category 3: Where it is reasonably certain that death was caused by both surgical and anesthesia factors. 83 4 Supracondylar 24 h after surgery Absence of preoperative management of risk amputation situations that could be compensated.
Invasive monitoring once the absence of invasive arterial line or perioperative transfusion could have contributed to a severe instability.
Better management of electrolyte disorder and reperfusion syndrome could have contributed to a better outcome.
Better hemodynamic control. Postoperative hypotension could have contributed to myocardial infarction and severe renal injury.
Protamine reaction could have contributed to irreversible hemodynamic instability. Anaphylactoid reaction should have been considered.
Prompt recognition of septic shock.
Category 2 - Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist. Better airway and extubation planning 63 3 Carotid aneurysm Immediately postThere is inappropriate choice of artificial airway or treatment operative failure to maintain or provide adequate protection of the airway. Where it is considered that the medical condition was a significant factor in anesthesia-related death. Better evaluation of the previous cardiac condition, 72 2 Resection of brain 14 days after Greater absolute or relative dosage according to The anesthetic induction could have been lighter. tumor surgery patient size or conditions. Where it is considered that the medical condition was a significant factor in anesthesia-related death
Age
Table 4 Description of the deaths related to anesthesia.
(continued on next page)
Patient had an arterial occlusion; he was a diabetic type 2 with acute renal failure and atrial fibrillation. Possible septic shock without prior compensation before surgery. Lower limb amputation was performed. He evolved to cardiac arrest, managed with resuscitation maneuvers. After, he had a progressive shock, refractory to treatment, and died 24 h after surgery. Advanced diabetes, hypertension with infected ulcer and ischemic pain at lower limb. Embolectomy was performed, but there was a reintervention after 12 h because of an arterial occlusion. During the procedure he presented refractory hemodynamic instability, with probable protamine reaction contribution. After a couple hours, he had a cardiac arrest and died at the ICU. Surgical urgent by-pass in a patient with poor functional capacity without cardiac or renal disease diagnosis. She remained hypotensive in postoperative period and evolved to a probable infarction. Also, she had surgical bleeding and rhabdomyolysis followed by severe renal injury and died 2 days after surgery. Smoker patient, with peripheral arterial disease followed by an acute ischemia in lower limb and an aortic bifemoral bypass. During surgery he evolved to a shock and pulseless arrest after releasing the aortic clamp. The cardiac rhythm was recovered after resuscitation maneuvers and he was referred to ICU with high norepinephrine dosage, metabolic acidosis and apparently an undiagnosed hyperkalemia. He evolved to a surgical bleeding, massive transfusion, and refractory shock. Patient with cholangiocarcinoma, cholangitis and acute renal failure in hemodialysis, submitted to a video cholecystectomy. He presented transoperative shock, with 2.500 ml of ascites, 800 ml bleeding. He was transferred to ICU intubated after surgery, with sensory deterioration and evolved to death in 24 h.
Smoker, with cerebral-vascular disease, submitted to a carotid surgery. After extubation he moved the four limbs, but upon awakening, he presented severe hypoxemia and in spite of the prompt airway maneuvers he died of irreversible cardiac arrest in asystolia. Patient was candidate to cerebral tumor resection. She was a diabetic, smoker, had controlled high blood pressure. After anesthetic induction with propofol, remifentanil and atracurio she had an atrial fibrillation and bradicardia that evolved to a cardiac arrest. It was reversed with adrenalin and thoracic compressions, but the surgical procedure was canceled and she died during the hospitalization due to pulmonary emboli 14 days after the cardiac arrest.
Case description
L.C. Stefani et al.
Journal of Clinical Anesthesia 49 (2018) 79–86
ASA
3
3
4
3
Age
77
77
76
60
Gastroenteroanastomosis
Carotid endarterectomy
Partial cholectomy
Open cholecystectomy
Procedure
Table 4 (continued)
5 days after surgery
24 h after surgery
3 days after surgery
7 days after surgery
Time from operation to death
Death resulted from inappropriate management or lack of active intervention by the anesthesiologist in the postoperative period. Where it is considered that the medical condition was a significant factor in anesthesia-related death.
No stratification or risk recognition in potentially severe patients. Where it is considered that medical condition was a significant factor in anesthesia-related death.
Inappropriate choice of anesthetic technique. Inappropriate management or lack of active intervention by the anesthetist in the postoperative period.
Death resulted from inappropriate management or lack of active intervention by the anesthetist in the postoperative period Where it is considered that the medical condition was a significant factor in anesthesia-related death.
Consensus subcategory
Diabetic type 2 with hypertension, amaurosis, cerebral-vascular disease, prostate cancer. He had an open cholecystectomy due to complicated cholecystitis. On the 1st postoperative day after the surgery, he evolved to a delirium hyperactive, on the 2nd day he had a convulsive crisis and atelectasis, on the 4th day he was confused and vomiting, with massive aspiration and was transferred to ICU. He had a cardiac arrest upon arrival at ICU. In spite of resuscitation maneuvers, he evolved to refractory shock and metabolic acidosis. He died 7 days after surgery. A female patient with colon neoplasia, hypertension, Class I heart failure and anemia underwent combined general and raquianesthesia to colon resection. She developed a postoperative moderate pain, atelectasis, and ileus. She underwent a massive aspiration in the 3th postoperative day and was transferred to the ICU after intubation at the ward, but evolved to a cardiac arrest, not responding to the resuscitation maneuvers. Patient with diabetes and chronic renal failure in hemodialysis presenting severe trivascular lesion in preoperative cardiac catheterization. He had a critical carotid stenosis bilateral and was submitted to carotid endarterectomy. He evolved to a postoperative myocardial infarction and was submitted to an urgent revascularization. He died 24 h from the first procedure. Advanced gastric neoplasm with peritoneal metastasis. He presented severe postoperative ileus from receiving a high opioid dosage and evolved to sepsis, dying in the 5th postoperative day.
Better analgesic strategy implementation and early sepsis recognition.
Prescription of multimodal analgesia, low opioid usage could have prevented ileus. Early sepsis recognition in the preoperative phase.
Cardiac revascularization procedure could have been done before the carotid surgery.
Multimodal analgesia with epidural catheter could have prevented atelectasis and ileus.
Case description
Possible prevention
L.C. Stefani et al.
Journal of Clinical Anesthesia 49 (2018) 79–86
85
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jclinane.2018.06.025.
[4]
Disclosures [5]
There are no financial conflicts of interest to disclose. Authors' contributions
[6]
All authors made a significant contribution to (a) the study concept and design, acquisition of data, or analysis and interpretation of data; (b) drafting/revising the manuscript for important intellectual content; and (c) approval of the final version to be published.
[7] [8]
[9]
Funding
[10]
This research was supported by grants and material support from the following Brazilian agencies: Postgraduate Program in Medical Sciences at the School of Medicine of the Federal University of Rio Grande do Sul (protocol number 14-0252). Postgraduate Research Group at the Hospital de Clínicas de Porto Alegre (FIPE-HCPA) (material support).
[11]
[12]
[13]
Conflicts of interest We declare we do not have any conflict of interest.
[14]
Congresses [15]
This research was presented as e-poster at European Congress of Anesthesia in Geneva, Jun 2017 [number: 16AP03-9]. [16]
References [1] Moonesinghe SR, Mythen MG, Grocott MPW. High-risk surgery: epidemiology and outcomes. Anesth Analg 2011;112:891–901. http://dx.doi.org/10.1213/ANE. 0b013e3181e1655b. [2] Moonesinghe SR, Mythen MG, Das P, Rowan KM, Grocott MP. Risk stratification tools for predicting morbidity and mortality in adult patients undergoing major surgery: qualitative systematic review. Anesthesiology 2013;119:959–81. http://dx. doi.org/10.1097/ALN.0b013e3182a4e94d. [3] Shah N, Hamilton M. Clinical review: can we predict which patients are at risk of
[18] [19] [20]
86
complications following surgery? Crit Care 2013;17:226. http://dx.doi.org/10. 1186/cc11904. Bainbridge D, Martin J, Arango M, Cheng D, Group EPCOR (EPiCOR). Perioperative and anaesthetic-related mortality in developed and developing countries: a systematic review and meta-analysis. Lancet 2012;380:1075–81. http://dx.doi.org/10. 1016/S0140-6736(12)60990-8. Pignaton W, Braz JRC, Kusano PS, Módolo MP, De Carvalho LR, Braz MG, et al. Perioperative and anesthesia-related mortality: an 8-year observational survey from a tertiary teaching hospital. Med (United States) 2016;95:1–6. http://dx.doi.org/10. 1097/MD.0000000000002208. Haller G, Laroche T, Clergue F. Morbidity in anaesthesia: today and tomorrow. Best Pract Res Clin Anaesthesiol 2011;25:123–32. http://dx.doi.org/10.1016/j.bpa. 2011.02.008. McNicol L. Safety of Anaesthesia; A Review of Anaesthesia-related Mortality Reporting in Australia and New Zealand 2009–2011. 2014. Glance LG, Lustik SJ, Hannan EL, Osler TM, Mukamel DB, Qian F, et al. The surgical mortality probability model. Ann Surg 2012;255:696–702. http://dx.doi.org/10. 1097/SLA.0b013e31824b45af. Fecho K, Lunney AT, Boysen PG, Rock P, Norfleet EA. Postoperative mortality after inpatient surgery: incidence and risk factors. Ther Clin Risk Manag 2008;4:681–8. Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029–34. http://dx.doi.org/10.1097/SLA.0b013e3181bef697. Pearse R, Moreno RP, Bauer P, Pelosi P, Metnitz P, Spies C, et al. Mortality after surgery in Europe: a 7 day cohort study. Lancet 2012;380:1059–65. http://dx.doi. org/10.1016/S0140-6736(12)61148-9. Guest GD, McLeod E, Perry WRG, Tangi V, Pedro J, Ponifasio P, et al. Collecting data for global surgical indicators: a collaborative approach in the Pacific Region. BMJ Glob Heal 2017;2:e000376http://dx.doi.org/10.1136/bmjgh-2017-000376. Neily J, Mills PD, Young-Xu Y, Carney BT, West P, Berger DH, et al. Association between implementation of a medical team training program and surgical mortality. JAMA J Am Med Assoc 2010;304:1693–700. http://dx.doi.org/10.1001/ jama.2010.1506. Baker R, Sullivan E, Camosso-Stefinovic J, Rashid A, Farooqi A, Blackledge H, et al. Making use of mortality data to improve quality and safety in general practice: a review of current approaches. Qual Saf Health Care 2007;16:84–9. http://dx.doi. org/10.1136/qshc.2006.019885. Theissen A, Fuz F, Bouregba M, Autran M, Beaussier M. A ten-year analysis of the reasons for death following ambulatory surgery: nine closed claims declared to the SHAM insurance. Anaesth Crit Care Pain Med 2018. http://dx.doi.org/10.1016/j. accpm.2018.03.001. Bainbridge D, Martin J, Arango M, Cheng D. Perioperative and anaesthetic-related mortality in developed and developing countries: a systematic review and metaanalysis. Lancet 2012;380:1075–81. http://dx.doi.org/10.1016/S0140-6736(12) 60990-8. Tiret L, Desmonts JM, Hatton F, Vourc'h G. Complications associated with anaesthesia - a prospective survey in France. Can Anaesth Soc J 1986;33:336–44. Surgery CG of AEWLRPeditor. Good Practice in Management of Emergency Surgery: A Literature Review. 2010. Lienhart A, Auroy Y, Péquignot F, Benhamou D, Warszawski J, Bovet M, et al. Survey of anesthesia-related mortality in France. Anesthesiology 2006;105:1087–97. http://dx.doi.org/10.1097/00000542-200612000-00008.
Contents lists available at ScienceDirect
Journal of Clinical Anesthesia journal homepage: www.elsevier.com/locate/jclinane
Original Contribution
Perioperative mortality related to anesthesia within 48 h and up to 30 days following surgery: A retrospective cohort study of 11,562 anesthetic procedures
T
⁎
Luciana C. Stefani, MD, PhDa,b, , Patricia W. Gamermann, MDb, Amanda Backof, MDb, Fernanda Guollo, MDb, Rafael M.J. Borgesc, Adriana Martin, MDd, Wolnei Caumo, MD, PhDa,b,e, Elaine A. Felix, MD, PhDa,b a
Department of Surgery, School of Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil Serviço de Anestesia e Medicina Perioperatória, Hospital de Clínicas de Porto Alegre (HCPA), Porto Alegre, Brazil School of Medicine, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil d Cleveland Clinic, Cleveland, USA e Pain and Neuromodulation Laboratory, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil b c
A R T I C LE I N FO
A B S T R A C T
Keywords: Perioperative mortality Anesthesia-related death Surgical risk In-hospital mortality
Study objective: Studying postoperative in-hospital mortality is crucial to the understanding of the perioperative process failures and to the implementation of strategies to improve patient outcomes. We intend to classify the causes of perioperative deaths up to 30 days after procedures requiring anesthesia and to evaluate the risk factors for early (48 h) or late (30 day) mortality. Design: Retrospective cohort study. Setting: A quaternary University Hospital from South Brazil. Patients: The information related to the perioperative care was collected from surgeries performed between January 2012 and December 2011. Interventions: None (observational study). Measurements: Three anesthesiologists classified the causes of deaths according to the ANZCA (Australian and New Zealand College of Anesthetists) classification, used in the report of Anesthesia-Related Mortality in Australia since 1985, which defines eight death categories. The risk factors for early or late death were analyzed in a regression model. Main results: 11.562 surgeries were performed, with a mortality incidence of 2.75% within 30 days (319 deaths). Most deaths were inevitable (50.7%), as they were related to advanced illnesses and would occur regardless of anesthetic or surgical procedures. The second most common cause was related to surgical complications (25%). The death rate having anesthesia as a likely contributor was 1.72:10.000 procedures, and as a potential contributor 7.78:10.000. These deaths occurred significantly earlier (< 48 h) when compared to deaths from other causes. Transoperative vasopressor, extremes of age and out-of-hour surgery were independent variables associated to early deaths. Conclusions: The study confirms that postoperative mortality in which anesthesia was involved occurred earlier in the perioperative period. In addition, it was revealed that this involvement of anesthesia as a morbidity contributor shows higher frequency when considering the anesthesiologist perioperative role, and when assessing the mortality in the long term (30 days).
1. Introduction Global perioperative risk is multifactorial. It depends on the interaction between anesthesia, patient's clinical conditions, and specific aspects of the surgery [1]. Preoperative stratification comprehends a ⁎
deep knowledge about the patient, the identification of precise risks, and to the adequate trans- and postoperative management. Postoperative morbidity and mortality can be assessed in different moments: in the transoperative period, in the immediate postoperative period, in 48 h, 30 days, or in one year or more following surgery.
Corresponding author at: Rua Ramiro Barcelos, 2350, 90035-003 Porto Alegre, RS, Brazil. E-mail addresses: [email protected] (L.C. Stefani), [email protected] (P.W. Gamermann), [email protected] (W. Caumo), [email protected] (E.A. Felix).
https://doi.org/10.1016/j.jclinane.2018.06.025 Received 20 February 2018; Received in revised form 2 June 2018; Accepted 8 June 2018 0952-8180/ © 2018 Elsevier Inc. All rights reserved.
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Fig. 1. Flowchart of the methodology used in the study.
Committee of Grupo de Pesquisa e Pós-Graduação from Hospital de Clínicas de Porto Alegre – Brazil (Chairperson Prof. Patricia Prolla) on 22 may 2014 (protocol number 14-0252.). The Hospital de Clínicas de Porto Alegre is a reference university quaternary care institution in southern Brazil. It holds 842 beds and performs around 20,000 surgeries a year, serving all age groups. The Anesthesiology and Perioperative Medicine Service comprises a permanent team of 83 anesthesiologists and 40 other resident doctors certified by the Brazilian Anesthesiology Society. The institution's information management system stores data from all surgical procedures and patients who underwent them. Using this system, we collected data from the deaths that occurred between anesthesia and up to 30 days post-surgery, from January 1st, 2012, to December 31st, 2013. The outcome was in-hospital death up to 30 days. When the patients overstay the 30-day period, the following days were not included in the analysis. Data were obtained by reviewing electronic and paper records (anesthesia files). The researchers were resident doctors trained to find information related to preoperative comorbidity, surgery and anesthesia, as well as postoperative complications, and details on the event of death. After data collection, three of the anesthesiologists
Statistics vary according to the time of assessment, but it is known that, usually, a small group of high-risk patients accounts for a greater part of the deaths and significantly longer hospitalizations [2, 3]. Therefore, identifying and signaling high-risk patients is essential to provide appropriate care, to implement strategies for preventing complications, and to precisely allocate resources. With the advancement of anesthesia and surgery techniques over the last 50 years, perioperative morbidity and mortality rates decreased due to improvements in safety, enhancements in monitoring techniques, the development and adoption of care protocols, and systematic measures for risk reduction [4–6]. Considering the evolution of anesthesia safety and its role in the perioperative medicine, it's reasonable to analyze the impact of anesthesia act in the postoperative mortality during a larger period of observation. This study aims to assess the causes of perioperative deaths within 30 days following procedures requiring anesthesia. We also intend to compare the risk factors associated to early (48 h) or late deaths (30 days). 2. Methods This retrospective cohort study had ethical approval by the Ethical 80
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
authors of this study (LS, AB, FG) met to discuss and classify the deaths and their probable causes. The ANZCA (Australian and New Zealand College of Anesthetists) [7] classification served as a background source in this process. This report of anesthesia-related mortality in Australia has been done since 1985. Its methodology defines eight death categories – the first three categories are considered deaths attributable to anesthesia, in the other three anesthesia played no part and deaths are related to surgical factors or other factors beyond the control of the physicians, and the last two categories are classified as unassessable deaths. For the deaths related somehow to anesthesia, there are 24 subcategories. The complete tool is detailed in Supplements 1 and 2. The strobe guideline for observational studies was followed and could be accessed in supplement.
Table 1 Characteristics of patients and procedures that resulted in death up to and after 48 h following surgery. Death < 48 h
2.1. Statistical analysis The results are presented as interquartile intervals, or absolute frequencies. Variables related to early and late deaths were compared using the chi-square test, Fisher's exact test, or Student's t-test for continuous data. A logistic regression model was used to identify which factors were independently associated with early deaths. Clinically significant variables or variables with p < 0.1 in the univariate analysis were included in the model. The significance level was set at p < 0.05 (bicaudal). Data analysis was performed using the Statistical Package for the Social Sciences software (SPPS), version 20.0 for MAC (SPSS Inc., Chicago, IL).
Initially, a total of 40.505 procedures were identified for the analyzed period. Their duplicates were discarded, as well as diagnostic and procedures under local anesthesia. The valid procedures were then arranged together and categorized into minor, intermediate, or major. These categories were based on previous study [8] and discussed among the institution's surgical specialists. The analysis took into consideration only one procedure per patient, choosing the largest one. Fig. 1 shows the study flowchart. A total of 11.562 surgeries were analyzed. The incidence of postoperative mortality within 48 h was 0.65%, while the incidence of postoperative mortality in patients hospitalized for up to 30 days was 2.75%. Most of these deaths were concentrated in patients ≥ ASA III (86.9%) submitted to non-elective surgeries, either urgent or emergent (55%). Table 1 presents patient characteristics, preoperative risk conditions, and information about the nature of the surgery, as well as surgery dimension, and the anesthetic technique used. It compares early (< 48 h) and late (> 48 h) deaths. The procedure with the highest rate of deaths was exploratory laparotomy (22.1% of total deaths), followed by colorectal resections (5.9%), and upper and lower limb amputations (5.6%). The exploratory laparotomy rate was higher than other procedures, 71 deaths in 461 procedures (15.4%). A logistic regression model identified the presence of extremes of age [OR 1,96 IC 1,04-3,69], transoperative vasopressors use [OR 3,36 IC 1,73–6,53) and out-of-hour surgery [OR 3,73 IC 1,36–10,18) as factors independently associated with early mortality in comparison with late mortality (Table 2). The most frequent causa mortis both in early and late deaths was sepsis (p = 0.06), followed by bleeding in early deaths (p < 0.01), and advanced neoplasm in early and late deaths (p = 0.42) (Fig. 2). A consensual qualitative analysis of the sample's death cases identified the causes of deaths and their relation with the anesthesia, the surgical procedure per se, the patient conditions or the combining of these factors. It has been shown that early deaths are more probably related to anesthesia than deaths over 48 h (p = 0.02) (Table 3). Most
p
n = 76
%
n = 243
%
Age (years) 0–1 2–15 16–35 36–50 51–64 65–79 ≥80
7 1 2 3 17 35 11
9.3 1.3 2.7 4.0 21.3 46.7 14.7
16 8 14 24 67 84 30
6.5 3.3 5.7 9.8 28.2 34.3 12.2
0.42 0.37 0.27 0.11 0.23 0.06 0.62
ASA I II III IV V
0 5 22 37 12
0 6.7 29 49.3 16
2 35 120 75 11
0.8 14.3 49.4 31 4.5
0.42 0.09 < 0.01 < 0.01 < 0.01
Degree of surgerya Minor Intermediary Major
3 17 56
4 22 73.3
33 77 133
13.5 31.4 55.1
0.02 0.13 0.01
Nature Urgent Elective
50 25
66.7 33.3
124 121
50.6 49
0.02
13
17.1
11
4.5
0.01
63
83
232
95.5
Perioperative risk situation Ischemic heart disease Renal insufficiency [cr > 2] Functional capacity < 4METs Cardiac insufficiency DM, insulin-dependent Hemodynamic instabilitya Bowel obstruction Preoperative anemia Preoperative sepsis Vasoactive drug Transfusion
8 10 65 10 6 30 12 20 17 50 28
10.7 13.3 86.7 13.3 8 40 16 26.7 22.7 66 37.3
27 33 190 29 29 43 35 83 33 68 31
11 13.5 77.6 11.8 11.8 17.6 14.3 33.9 13.5 27.8 12.7
0.92 1.0 0.09 0.76 0.37 < 0.01 0.69 0.23 0.06 < 0.01 < 0.01
Anesthesia type General anesthesia Regional anesthesia Sedation Combined anesthesia
72 2 0 0
97.3 2.7 0 0
199 27 3 16
81.2 11 1.2 6.5
< 0.01 0.03 0.32 0.02
Schedule Out-of-hour surgery [after midnight] Surgery before midnight
3. Results
Death > 48 h
a
Uncontrolled hypotension, or hypoperfusion signs.
deaths were considered inevitable (50.9%), meaning that they would occur irrespective of the anesthetic or surgical act, but due to the patient's unfavorable conditions. The second category identified were deaths related to surgery in which anesthesia was not a contributor (26,1%). Two cases were classified as related to anesthesia or other factors controlled by the anesthetist (category 2: Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist). These two deaths were associated with cardiac arrest at anesthetic induction in the first case and with immediate postoperative hypoxemia and cardiac arrest in the second case. Nine cases were categorized as having been caused by both surgical or anesthetic acts. Among these deaths partially related to anesthesia (category 3), the main contributing factors were hemorrhagic shock (38.5%), followed by sepsis (23%), thromboembolic phenomena - such as acute myocardial infarction (15.38%), and complications associated to hypoxemia - such as aspiration and pneumonia (15.38%). The description of all cases possible related to anesthesia is in Table 4.
81
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Table 2 Logistic regression of risk factors for deaths up to 48 h vs. after 48 h. C coeficiente (B)
Standard error
p value
Odds ratio
95% IC
Variables related to patients Extreme age ASA > 3 Low functional capacityb Pre or transoperative transfusion Transoperative continuous vasoactive drugs Hemodynamic instability after surgeryc
0,67 0,64 −0,45 0,70 1,21 0,48
0,32 0,51 0,41 0,38 0,33 0,34
0,03 0,35 0,27 0,067 0,00 0,15
1,96 1,90 0,63 2,02 3,36 1,62
1,04–3,69 0,64–5,6 0,28–1,42 0,95–4,31 1,73–6,53 0,82–3,18
Variables related to surgery Surgery duration > 150 min Emergency operation Out-of-hour surgerya Anesthesia type
0,28 0,24 1,31 1,47
0,33 0,34 0,51 0,77
0,40 0,46 0,01 0,058
1,32 1,27 3,73 4,35
0,68–2,56 0,65–2,4 1,36–10,18 0,95–19,91
Odds ratios were calculated with the use of multivariable logistic regression and they are for the group of patients that died before 48 h as compared with those that died after 48 h from the surgery. An odds ratio of > 1 represents a greater likelihood of death in < 48 h after surgery. Listed are variables with p < 0,1 in the multivariable model. Risk factors: age (0–1 or +65 years vs. 2–64 years), ASA PS (3, 4, 5 vs. 1, 2), low functional capacity/transfusion, vasoactive drugs, hemodynamic instability (yes vs. no), Anesthesia type (general vs. regional or combined or sedation). a Unsafe working hours (after midnight to seven). b Capacity < 4 METs [metabolic milliequivalent]. c Intensive care resource recovery or hemodynamic monitoring after surgery.
Our study has two main contributions: first, we looked deeper into the anesthesia influence on postoperative deaths up to 30 days reviewing all cases in a consensual manner and classifying their probable causes. Among the identifiable causes, the patients' conditions were the most relevant contributing factor, resulting in inevitable deaths that would occur regardless of surgery or anesthesia (50%). The second most common cause identified was related to complications from the surgical procedure (25%). Although, anesthesia was almost certainly involved in 1.72:10.000 procedures, and was a potential contributor in 7.78:10.000 cases. This is significantly higher than the most recent data, published from an Australian audit, which was around 0.17:10.000 [7]. However, this is a national audit from five Australian states, and as it is stated on the conclusion, they do not have available methods to verify whether all anesthesia and/or sedation deaths are reported under that provision. Furthermore, they evaluate only deaths up to 24 h after surgery, and we evaluate deaths up to 30 days after surgery. Otherwise, a recently audit from deaths observed within the 30-day period following anesthesia declared to the SHAM insurance (Société Hospitalière d'Assurance Mutuelle-Paris) over the last 10 years found concordant levels to our cohort. They showed an incidence of 0.88 per 10.000 deaths related to anesthesia in ambulatory patients [15].
4. Discussion This study identified a perioperative mortality rate of 0.64% within 48 h after the procedure and 2.74% in patients hospitalized for up to 30 days following surgery. In general, most deaths occurred later after surgery, in critical patients (ASA ≥ III) submitted to urgent surgeries and were associated with complications such as sepsis and advanced neoplastic disease. Our results are consistent with postoperative death occurrences in hospitalized patients in developed countries, as shown by Fecho et al. in 2008 [9] who has presented a similar death rate: 0.57% within 48 h and 2.1% up to 30 days. Ghaferi et al. [10], however, assessed mortality rates in different American hospitals in 2009 and found values that ranged from 3.5% to 6.9%. This is consistent with a recent European study [11] that compared various countries for perioperative mortality of inpatients, pointing out the negative correlation between mortality and the Human Development Index of each country. Despite recent advances in quality improvements and safety of the surgical patient [12, 13] there is scarce recent research focusing on anesthesia safety and on the role of the anesthetist as a perioperative physician capable to optimize outcomes. Therefore, postoperative mortality in 48 h and 30 days can be used to evaluate the anesthesia performance beyond the transoperative period [14].
Fig. 2. Causa mortis associated with deaths < 48 h or > 48 h. Bars followed by (*) present significant differences.
82
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Table 3 Incidence of deaths and categories analyzed in the perioperative period of 11,562 procedures. Death category according to the ANZCA classification
Death attributable to anesthesia I) Where it is reasonably certain that death was caused by the anesthesia or other factors under the control of the anesthetist II) Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist. III) Where it is reasonably certain that death was caused by both surgical and anesthesia factors. Death not attributable to anesthesia IV) Death where the administration of the anesthesia is not contributory and surgical or other factors are implicated. V) Inevitable death, which would have occurred irrespective of anesthesia or surgical procedures. VI) Incidental death which could not reasonably be expected to have been foreseen by those looking after the patient, was not related to the indication for surgery and was not due to factors under the control of the anesthetist or surgeon. VII) Those that cannot be assessed despite considerable data but where the information is conflicting or key data are missing. VIII) Cases that cannot be assessed because of inadequate data.
Death < 48 h n = 76
Death > 48 h n = 243
Total (%)
p
– 1
– 1
– 2 (0.6%)
– 0.37
5
4
8 (2,5%)
0.02
25 41 1
58 122 51
0.13 0.55 0.00
2 1
6 1
83 (26,1%) 162 (50,9%) 52 (16.4%) 8 (2.5%) 2 (0.6%)
0.92 0.37
adequate management of the surgical schedule, focusing in avoiding out-of-hour's procedures that could have been done in diurnal shift. The anesthetists could serve as a median point between nurses and surgeons in this patient priority decision-making process [11, 19]. The major strength of our study is the exhaustive revision of the pre, trans e postoperative care period in order to identify possible death risk factors linked somehow to the anesthesia act. Preoperatively suboptimal care related to inadequate patient evaluation or suboptimal management of postoperative complications occurred in 63% of our deaths related to anesthesia. Thus, to improve the anesthesia safety, institutions should prioritized the perioperative assistance as much as the intraoperative period [6, 20]. The information related to surgical procedures and the general patients' data were collected from the hospital's information management system and the researchers individually assessed the medical files. The long period of assessment (30 days) allowed us to evaluate late complications and identify their possible associated causes. Using a tool to classify deaths into categories based on a consensus among three anesthetists reduces the subjectivity of the classification, which could be a problem in studies based on a single reviewer. Nevertheless, it is not possible to evaluate the direct influence of teamwork, failure in communication, or leadership problems during the intercurrences in a retrospective study. These non-technical skills are always present on the root cause analysis of any adverse event and should not be minimized. Methodological limitations include the observational design, which is exploratory and dependent on the quality of the data registry. This means that transoperative data rely on the commitment of each anesthetist to the registry process. Additionally, the data is originated from a single institution and may not reflect other scenarios.
We could point out the lack of deaths caused reasonably certainly by anesthesia in our cohort. Although, in two cases, the anesthesia was involved in negative outcomes once the postoperative period handling and the incorrect judgment of patients' particular clinical conditions led to instabilities during the induction and extubation phases. We also found that other deaths somehow related to, but not caused by anesthesia, had influence on both surgical and anesthetic factors. Probably, these cases wouldn't have been even recognized if other less rigid classification tool was adopted, since bleeding, shock, sepsis, and severe ileus were likely considered common surgical complications without anesthesia's involvement. The detailed cases revision revealed that the perioperative commitment of the anesthesiologist seems suboptimal, since the lack of preoperative compensation or the development of postoperative complications could have been avoided, in some of the cases, by reducing the perioperative care fragmentation. The second contribution of our study is the confirmation that the vast majority of deaths with any connection to the anesthetist's performance were of early occurrence (p < 0.01). We found that 7:76 deaths (9,3%) in < 48 h were related to anesthesia; compared to 5:246 from late deaths (2,4%, p ≤ 0,01). The early deaths often reflect the comorbidities and physiological disorders of the patient, the quality and complexity of surgical care, or the risks of anesthesia. However, it cannot be used to compare one site, facility, or country with another without appropriate, validated, and time-consuming risk adjustment [16]. Independently of the cause, the risk factors significantly associated to early deaths were extremes of age (OR 1,96 IC 1,04–3,69), transoperative continuous vasoactive drugs (OR 3,36 IC 1,73–6,53) and out-of-hour surgeries (OR 3,73 IC 1,36–10,28). Extremes of age are associated to physiological fragility and have been identified as a risk factor for perioperative adverse events [10,18,19]. In our study, infants (0–1 year old) and older people (65+ years old) had higher rates of early postoperative mortality than people ranging 2–64 years old. The procedures under these populations should have an extra attention from the caregivers and be prioritized in the list of emergent surgeries. Moreover, transoperative vasoactive drug use reflects the patient severity or the complexity of surgical procedure and it could be associated to any kind of shock, therefore it could be considered an epiphenomenon. Also, performing a surgery during the night (from19:00 h to 07:00 h) had clear impact on outcomes: prolonged hours of work and night shifts bring with them the well documented risks to the performance of the surgical and anesthetic team, and therefore, the quality of patient care provided. The potentially serious consequence of fatigue increases the risk of serious errors that can lead to death or serious morbidity. The operational efficiency of a surgical center involves the
5. Conclusions This study showed that the majority of late postoperative deaths were inevitable (50.7%), as they were related to advanced illnesses and would happen regardless of the anesthetic or surgical acts. Differently from traditional studies on deaths related to anesthesia, we looked deeper and longer into the perioperative path, and found that anesthesia is almost certainly involved in 1.72:10.000 procedures, and was a potential contributor in 10.37:10.000 cases. Furthermore, death cases in which the anesthetic contribution was recognized were significantly premature (< 48 h) when compared to other investigated causes. This study reinforced the need to implement specific multiprofessional teams dedicated to the pre, trans and postoperative care of the high-risk patients, focusing on risk management and reducing the failure to rescue during the perioperative journey.
83
ASA
Procedure
Time from operation to death
Consensus subcategory
Possible prevention
84
3
3
4
75
62
4
56
73
Cholecystectomy
Bypass femoropopliteus
Bypass femoropopliteus
Bypass with stent
24 h after surgery
24 h after surgery
48 h
< 24 h after surgery
Failure to comply with minimum monitoring standards or lack of advanced monitoring when indicated
Death resulted from inappropriate management or lack of active intervention by the anesthetist in the postoperative period. Patient clinical conditions were significant for the occurrence of death related to anesthesia Unsuitable handling of unexpected occurrences during anesthesia
Unsuitable handling of unexpected occurrences during anesthesia. Patient clinical conditions were significant for the occurrence of death related to anesthesia
Category 3: Where it is reasonably certain that death was caused by both surgical and anesthesia factors. 83 4 Supracondylar 24 h after surgery Absence of preoperative management of risk amputation situations that could be compensated.
Invasive monitoring once the absence of invasive arterial line or perioperative transfusion could have contributed to a severe instability.
Better management of electrolyte disorder and reperfusion syndrome could have contributed to a better outcome.
Better hemodynamic control. Postoperative hypotension could have contributed to myocardial infarction and severe renal injury.
Protamine reaction could have contributed to irreversible hemodynamic instability. Anaphylactoid reaction should have been considered.
Prompt recognition of septic shock.
Category 2 - Where there is some doubt whether death was entirely attributable to the anesthesia or other factors under the control of the anesthetist. Better airway and extubation planning 63 3 Carotid aneurysm Immediately postThere is inappropriate choice of artificial airway or treatment operative failure to maintain or provide adequate protection of the airway. Where it is considered that the medical condition was a significant factor in anesthesia-related death. Better evaluation of the previous cardiac condition, 72 2 Resection of brain 14 days after Greater absolute or relative dosage according to The anesthetic induction could have been lighter. tumor surgery patient size or conditions. Where it is considered that the medical condition was a significant factor in anesthesia-related death
Age
Table 4 Description of the deaths related to anesthesia.
(continued on next page)
Patient had an arterial occlusion; he was a diabetic type 2 with acute renal failure and atrial fibrillation. Possible septic shock without prior compensation before surgery. Lower limb amputation was performed. He evolved to cardiac arrest, managed with resuscitation maneuvers. After, he had a progressive shock, refractory to treatment, and died 24 h after surgery. Advanced diabetes, hypertension with infected ulcer and ischemic pain at lower limb. Embolectomy was performed, but there was a reintervention after 12 h because of an arterial occlusion. During the procedure he presented refractory hemodynamic instability, with probable protamine reaction contribution. After a couple hours, he had a cardiac arrest and died at the ICU. Surgical urgent by-pass in a patient with poor functional capacity without cardiac or renal disease diagnosis. She remained hypotensive in postoperative period and evolved to a probable infarction. Also, she had surgical bleeding and rhabdomyolysis followed by severe renal injury and died 2 days after surgery. Smoker patient, with peripheral arterial disease followed by an acute ischemia in lower limb and an aortic bifemoral bypass. During surgery he evolved to a shock and pulseless arrest after releasing the aortic clamp. The cardiac rhythm was recovered after resuscitation maneuvers and he was referred to ICU with high norepinephrine dosage, metabolic acidosis and apparently an undiagnosed hyperkalemia. He evolved to a surgical bleeding, massive transfusion, and refractory shock. Patient with cholangiocarcinoma, cholangitis and acute renal failure in hemodialysis, submitted to a video cholecystectomy. He presented transoperative shock, with 2.500 ml of ascites, 800 ml bleeding. He was transferred to ICU intubated after surgery, with sensory deterioration and evolved to death in 24 h.
Smoker, with cerebral-vascular disease, submitted to a carotid surgery. After extubation he moved the four limbs, but upon awakening, he presented severe hypoxemia and in spite of the prompt airway maneuvers he died of irreversible cardiac arrest in asystolia. Patient was candidate to cerebral tumor resection. She was a diabetic, smoker, had controlled high blood pressure. After anesthetic induction with propofol, remifentanil and atracurio she had an atrial fibrillation and bradicardia that evolved to a cardiac arrest. It was reversed with adrenalin and thoracic compressions, but the surgical procedure was canceled and she died during the hospitalization due to pulmonary emboli 14 days after the cardiac arrest.
Case description
L.C. Stefani et al.
Journal of Clinical Anesthesia 49 (2018) 79–86
ASA
3
3
4
3
Age
77
77
76
60
Gastroenteroanastomosis
Carotid endarterectomy
Partial cholectomy
Open cholecystectomy
Procedure
Table 4 (continued)
5 days after surgery
24 h after surgery
3 days after surgery
7 days after surgery
Time from operation to death
Death resulted from inappropriate management or lack of active intervention by the anesthesiologist in the postoperative period. Where it is considered that the medical condition was a significant factor in anesthesia-related death.
No stratification or risk recognition in potentially severe patients. Where it is considered that medical condition was a significant factor in anesthesia-related death.
Inappropriate choice of anesthetic technique. Inappropriate management or lack of active intervention by the anesthetist in the postoperative period.
Death resulted from inappropriate management or lack of active intervention by the anesthetist in the postoperative period Where it is considered that the medical condition was a significant factor in anesthesia-related death.
Consensus subcategory
Diabetic type 2 with hypertension, amaurosis, cerebral-vascular disease, prostate cancer. He had an open cholecystectomy due to complicated cholecystitis. On the 1st postoperative day after the surgery, he evolved to a delirium hyperactive, on the 2nd day he had a convulsive crisis and atelectasis, on the 4th day he was confused and vomiting, with massive aspiration and was transferred to ICU. He had a cardiac arrest upon arrival at ICU. In spite of resuscitation maneuvers, he evolved to refractory shock and metabolic acidosis. He died 7 days after surgery. A female patient with colon neoplasia, hypertension, Class I heart failure and anemia underwent combined general and raquianesthesia to colon resection. She developed a postoperative moderate pain, atelectasis, and ileus. She underwent a massive aspiration in the 3th postoperative day and was transferred to the ICU after intubation at the ward, but evolved to a cardiac arrest, not responding to the resuscitation maneuvers. Patient with diabetes and chronic renal failure in hemodialysis presenting severe trivascular lesion in preoperative cardiac catheterization. He had a critical carotid stenosis bilateral and was submitted to carotid endarterectomy. He evolved to a postoperative myocardial infarction and was submitted to an urgent revascularization. He died 24 h from the first procedure. Advanced gastric neoplasm with peritoneal metastasis. He presented severe postoperative ileus from receiving a high opioid dosage and evolved to sepsis, dying in the 5th postoperative day.
Better analgesic strategy implementation and early sepsis recognition.
Prescription of multimodal analgesia, low opioid usage could have prevented ileus. Early sepsis recognition in the preoperative phase.
Cardiac revascularization procedure could have been done before the carotid surgery.
Multimodal analgesia with epidural catheter could have prevented atelectasis and ileus.
Case description
Possible prevention
L.C. Stefani et al.
Journal of Clinical Anesthesia 49 (2018) 79–86
85
Journal of Clinical Anesthesia 49 (2018) 79–86
L.C. Stefani et al.
Supplementary data to this article can be found online at https:// doi.org/10.1016/j.jclinane.2018.06.025.
[4]
Disclosures [5]
There are no financial conflicts of interest to disclose. Authors' contributions
[6]
All authors made a significant contribution to (a) the study concept and design, acquisition of data, or analysis and interpretation of data; (b) drafting/revising the manuscript for important intellectual content; and (c) approval of the final version to be published.
[7] [8]
[9]
Funding
[10]
This research was supported by grants and material support from the following Brazilian agencies: Postgraduate Program in Medical Sciences at the School of Medicine of the Federal University of Rio Grande do Sul (protocol number 14-0252). Postgraduate Research Group at the Hospital de Clínicas de Porto Alegre (FIPE-HCPA) (material support).
[11]
[12]
[13]
Conflicts of interest We declare we do not have any conflict of interest.
[14]
Congresses [15]
This research was presented as e-poster at European Congress of Anesthesia in Geneva, Jun 2017 [number: 16AP03-9]. [16]
References [1] Moonesinghe SR, Mythen MG, Grocott MPW. High-risk surgery: epidemiology and outcomes. Anesth Analg 2011;112:891–901. http://dx.doi.org/10.1213/ANE. 0b013e3181e1655b. [2] Moonesinghe SR, Mythen MG, Das P, Rowan KM, Grocott MP. Risk stratification tools for predicting morbidity and mortality in adult patients undergoing major surgery: qualitative systematic review. Anesthesiology 2013;119:959–81. http://dx. doi.org/10.1097/ALN.0b013e3182a4e94d. [3] Shah N, Hamilton M. Clinical review: can we predict which patients are at risk of
[18] [19] [20]
86
complications following surgery? Crit Care 2013;17:226. http://dx.doi.org/10. 1186/cc11904. Bainbridge D, Martin J, Arango M, Cheng D, Group EPCOR (EPiCOR). Perioperative and anaesthetic-related mortality in developed and developing countries: a systematic review and meta-analysis. Lancet 2012;380:1075–81. http://dx.doi.org/10. 1016/S0140-6736(12)60990-8. Pignaton W, Braz JRC, Kusano PS, Módolo MP, De Carvalho LR, Braz MG, et al. Perioperative and anesthesia-related mortality: an 8-year observational survey from a tertiary teaching hospital. Med (United States) 2016;95:1–6. http://dx.doi.org/10. 1097/MD.0000000000002208. Haller G, Laroche T, Clergue F. Morbidity in anaesthesia: today and tomorrow. Best Pract Res Clin Anaesthesiol 2011;25:123–32. http://dx.doi.org/10.1016/j.bpa. 2011.02.008. McNicol L. Safety of Anaesthesia; A Review of Anaesthesia-related Mortality Reporting in Australia and New Zealand 2009–2011. 2014. Glance LG, Lustik SJ, Hannan EL, Osler TM, Mukamel DB, Qian F, et al. The surgical mortality probability model. Ann Surg 2012;255:696–702. http://dx.doi.org/10. 1097/SLA.0b013e31824b45af. Fecho K, Lunney AT, Boysen PG, Rock P, Norfleet EA. Postoperative mortality after inpatient surgery: incidence and risk factors. Ther Clin Risk Manag 2008;4:681–8. Ghaferi AA, Birkmeyer JD, Dimick JB. Complications, failure to rescue, and mortality with major inpatient surgery in medicare patients. Ann Surg 2009;250:1029–34. http://dx.doi.org/10.1097/SLA.0b013e3181bef697. Pearse R, Moreno RP, Bauer P, Pelosi P, Metnitz P, Spies C, et al. Mortality after surgery in Europe: a 7 day cohort study. Lancet 2012;380:1059–65. http://dx.doi. org/10.1016/S0140-6736(12)61148-9. Guest GD, McLeod E, Perry WRG, Tangi V, Pedro J, Ponifasio P, et al. Collecting data for global surgical indicators: a collaborative approach in the Pacific Region. BMJ Glob Heal 2017;2:e000376http://dx.doi.org/10.1136/bmjgh-2017-000376. Neily J, Mills PD, Young-Xu Y, Carney BT, West P, Berger DH, et al. Association between implementation of a medical team training program and surgical mortality. JAMA J Am Med Assoc 2010;304:1693–700. http://dx.doi.org/10.1001/ jama.2010.1506. Baker R, Sullivan E, Camosso-Stefinovic J, Rashid A, Farooqi A, Blackledge H, et al. Making use of mortality data to improve quality and safety in general practice: a review of current approaches. Qual Saf Health Care 2007;16:84–9. http://dx.doi. org/10.1136/qshc.2006.019885. Theissen A, Fuz F, Bouregba M, Autran M, Beaussier M. A ten-year analysis of the reasons for death following ambulatory surgery: nine closed claims declared to the SHAM insurance. Anaesth Crit Care Pain Med 2018. http://dx.doi.org/10.1016/j. accpm.2018.03.001. Bainbridge D, Martin J, Arango M, Cheng D. Perioperative and anaesthetic-related mortality in developed and developing countries: a systematic review and metaanalysis. Lancet 2012;380:1075–81. http://dx.doi.org/10.1016/S0140-6736(12) 60990-8. Tiret L, Desmonts JM, Hatton F, Vourc'h G. Complications associated with anaesthesia - a prospective survey in France. Can Anaesth Soc J 1986;33:336–44. Surgery CG of AEWLRPeditor. Good Practice in Management of Emergency Surgery: A Literature Review. 2010. Lienhart A, Auroy Y, Péquignot F, Benhamou D, Warszawski J, Bovet M, et al. Survey of anesthesia-related mortality in France. Anesthesiology 2006;105:1087–97. http://dx.doi.org/10.1097/00000542-200612000-00008.