Abstract
Background:
There is a paucity of data on older adults (age ≥65 years) undergoing surgery who had an inpatient DNR order, and the association between timing of DNR order and outcomes.
Methods:
This was a retrospective analysis of 1976 older adults in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) geriatric-specific database (2014-18). Patients were stratified by institution of a DNR order during their surgical admission (“new-DNR” vs “no-DNR”), and matched by age (± 3years), frailty score (range: 0-1), and procedure. The main outcome of interest was occurrence of death or hospice transition (DoH) ≤ 30 postoperative days; this was analyzed using bivariate and multivariable methods.
Results:
One in 36 older adults had a new-DNR order. After matching, there were 988 new-DNR and 988 no-DNR patients. Median age and frailty score were 82yrs and 0.2, respectively. Most underwent orthopedic (47.6%), general (37.6%), and vascular procedures (8.4%). Overall DoH rate ≤ 30 days was 44.4% for new-DNR vs 4.0% for no-DNR patients (p<0.001). DoH rate for patients who had DNR orders placed in the preoperative, day of surgery, and postoperative setting was 16.7%, 23.3% and 64.6%, respectively (p<0.001). In multivariable analysis, compared to no-DNR patients, those with a new-DNR order had a 28-fold higher adjusted odds of DoH (odds ratio [OR] 28.1, 95% confidence interval: 13.0-60.1, p<0.001) however, odds were 10-fold lower if the DNR order was placed preoperatively (OR: 5.8 p=0.003) vs. postoperatively (OR: 52.9, p<0.001). Traditional markers of poor postoperative outcomes such as ASA class and emergency surgery were not independently associated with DoH.
Conclusions:
An inpatient DNR order was associated with risk of DoH independent of traditional markers of poor surgical outcomes. Further research is needed to understand factors leading to a DNR order that may aid early recognition of high-risk older adults undergoing surgery.
Keywords: surgery, geriatrics, DNR, outcomes, hospice, frailty
INTRODUCTION
Older adults (≥65years) represent 15.2% of the US population but approximately 40% of inpatient operations performed in the country [1]. Compared to younger patients, older adults experience higher rates of postoperative complications, functional decline, and mortality [2-4]. This is in part explained by geriatric-specific syndromes such as frailty [5-12]. Older adults also account for up to 80% of surgical patients with a Do-Not-Resuscitate (DNR) order which precludes cardiopulmonary resuscitation (CPR) in a patient experiencing cardiopulmonary arrest – itself a rare event in the perioperative period [13-15]. Studies have consistently shown an association between a DNR order and morbidity/mortality in medical and surgical patients [16-18].
Previous publications on the association of DNR orders with surgical outcomes have focused on patients with a pre-existing DNR order (placed prior to hospitalization) [13-14, 18-24]. There is a paucity of research on older adults who present to the hospital without a DNR order (full code status) and then have DNR order instituted during their surgical admission. Information about the characteristics and outcomes of these patients could provide insight into the care dynamics of extremely high-risk patients. On one hand, older adults who have a code status change in the immediate preoperative setting may represent a group of patients with high acuity of surgical illness and perceived low likelihood of survival following operative intervention. On the other hand, those who have a DNR order placed postoperatively may represent a separate category of patients whose clinical status worsened considerably after surgery. Both groups represent an opportunity to assess and improve the provision of surgical care to high-risk older adults undergoing surgery.
The ACS-NSQIP geriatric surgery pilot database provides a unique opportunity to address the current paucity of data for older surgical patients who have a code status change from full code to DNR during their surgical admission. The aim of this study was to characterize older surgical patients who had a DNR order placed during their hospitalization and their outcomes, and the association between the timing of the DNR order and outcomes.
METHODS
Data Source
This was a retrospective analysis of data captured in the American College of Surgeons - National Surgical Quality Improvement Program (ACS-NSQIP) geriatric surgery pilot database and participant user files (2014-2018). The database includes geriatric-specific information such as cognitive competency, fall history, and postoperative delirium for more than 40,000 older adults (age ≥65 years) who underwent surgery at 30 participating hospitals in the United States (see Supplemental material; Supplemental Table S1). Using the unique identifier for each case encoded in the databases, variables in the geriatric database were linked to the ACS-NSQIP main files for the corresponding years.
Patients
In the ACS-NSQIP geriatric surgery database, patients with a DNR order are defined as those with “a new DNR order (placed) during the hospitalization that was not present on admission”. This variable is new and differs from the DNR variable included in the ACS-NSQIP participant user files for years 2005 to 2012 (now discontinued), which represented patients who had a pre-existing DNR placed prior to hospitalization. ACS-NSQIP does not capture data on DNR order rescindment and reinstatement, as well as inter-specialty patient transfers.
Using an optimal matching algorithm, patients who had a DNR order instituted during their admission (“new-DNR”) were matched 1:1 to those without a DNR order (“no-DNR”) by age (± 3 years), exact modified frailty index (mFI) score, and exact procedure (using Current Procedural Terminology [CPT] codes); for instance, a 79 year old new-DNR patient with a frailty score of 0.6 undergoing a laparoscopic right hemicolectomy with a CPT code 44205 was matched to a 79 (±3) year old no-DNR patient with a frailty score of 0.6 undergoing a right hemicolectomy with a CPT code 44205. The validated 5-factor modified frailty score (mFI-5) comprising functional dependence, diabetes, COPD, CHF, and hypertension requiring medication, was computed as described by Subramanian et al [25-28] with scores ranging from 0 (minimum) to 1 (maximum).
Phase of care in which a new-DNR order was placed was computed based on variables included in the ACS-NSQIP database as follows: patients with a preoperative new-DNR order had their order instituted in the ER; patients whose number of “days from hospital admission to surgery” was equal to the number of “days from operation to DNR order” were defined as those whose DNR order was placed on the day of surgery. Patients with a postoperative order were those whose “days from operation to DNR order” was greater than “days from hospital admission to surgery”. Patient location (ER, ICU, and acute care ward) at the time the DNR order was placed was also examined.
Baseline characteristics
Demographic factors included age, gender, race, transfer status (admitted from another care facility or from home), functional status prior to surgery (independent, partially or totally dependent), which is based on the patient’s ability to perform activities of daily living (ADLs) in the 30 days prior to surgery; history of a fall within year of surgery, having advanced care planning prior to surgery (i.e. documentation of a healthcare proxy, living will or advanced directives), history of dementia, and cognitive competency to sign the surgical informed consent form at the time of surgery.
Clinical characteristics included body mass index, American Society of Anesthesiologists (ASA) classification, preoperative laboratory values (white blood cell count, hematocrit, INR, albumin, and creatinine); surgeon specialty, and urgency of the operation: elective vs. non-elective, and technical complexity of the operation as measured in work Relative Value Units. Preoperative laboratory values were modeled as categorical variables using conventional laboratory ranges. Comorbidities included disseminated cancer, coagulopathy, preoperative sepsis, ascites, acute renal failure, dialysis dependence, and preoperative ventilator dependence.
Outcomes
The primary outcome of interest was postoperative death or hospice transition (DoH); this was a composite of the occurrence of death or placement of a hospice order ≤30 postoperative days. A patient who was transitioned to hospice and then died ≤30 postoperative days was counted as one occurrence of the primary outcome.
Complications were analyzed as dichotomous variables; definitions of complications and other variables captured in ACS-NSQIP are available at https://www.facs.org/-/media/files/quality-programs/nsqip/nsqip_puf_userguide_2018.ashx [29]. In addition to complications captured in the ACS-NSQIP participant user file, the overall complication rate included the 18 complications (other than death, reoperation, and readmission) captured in ACS-NSQIP participant user file (see Supplemental Material; Supplemental Table S2) as well as postoperative delirium and pressure ulcer development (or progression) during the surgical admission; the latter 2 complications are reported only in the geriatric surgery database. The most common first complication and the median time of its occurrence from the day of surgery was analyzed.
Statistical analysis
Bivariate analyses comparing characteristics and outcomes of new- vs no-DNR patients were performed using two-tailed χ2 and Fisher analysis for categorical variables and 2-sided t- and Wilcoxon rank sum test for continuous variables. All p values <0.05 were considered significant. Since the study was based on matched data, conditional logistic regression was used to investigate the association between a new-DNR order and DoH transition, while controlling for confounders. Adjusted odds ratios with 95% confidence intervals (CI) were calculated. Because a conditional logistic regression model precludes the calculation of a concordance (c)-statistic (which measures the goodness-of-fit of a model), a second regression model including the exact risk factors included in the conditional regression model was run without conditionality, thereby allowing c-statistic computation.
Data analyses and management were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC). The ACS-NSQIP participant user and geriatric surgery pilot data files are entirely de-identified; this study was granted exemption by the Duke University Health System’s Institutional Review Board.
RESULTS:
Overall sample
Among 41481 older adults in the ACS-NSQIP geriatric surgery database, 1158 had a new-DNR order placed during their surgical admission. Therefore, the overall new-DNR order rate was 2.8% or 1 order per 36 older adults. Institution of a new-DNR order increased with age: 1.2% (1 in 83) for patients aged 64-75yrs, 3.0% (1 in 33) for those 76-84yrs, and 10.3% (1 in 9) for patients ≥85yrs. Upon matching by age ≤3yrs, exact mFI-5 score, and exact primary CPT code, 988 patients with a new-DNR order and 988 without a DNR order (85.3% match rate) were included in the study.
Characteristics of new-DNR vs. no-DNR patients
For the matched cohort of 1976 patients (988 new- and 988 no-DNR patients), overall median age was 82yrs (interquartile range [IQR]: 75-88 years). Most patients were white (73.1%), female (60.9%) and admitted from home (78.1%), and approximately 82.5% had a modified frailty score ≤0.2, which indicates low frailty. The median number of days from surgery to placement of a new-DNR order was 3 days (IQR: 1-10 days). Approximately 7.3%, 26.2%, and 30.6% of new-DNR patients had their DNR order instituted in the ER, ICU, and acute care ward, respectively.
Patients with a new-DNR order were more likely to have dementia, high ASA, functional impairment, and undergo non-elective procedures (Table 1; see also Supplemental Material; Supplemental Table S3). The most common operations were orthopedic followed by general surgical procedures. By specialty, the most common operations were femur fracture repair (62.2%) for orthopedic surgery patients, colectomy with/without colostomy (52.0%) for patients undergoing general surgery procedures, and aortic procedures (32.2%) for vascular surgery patients. There was no significant difference in overall operative time (mean: 111.5 ± 94.4 vs 117.8 ± 90.4 minutes for new- vs. no-DNR, p=0.125), and work Relative Value Units, a standard measure of the technical complexity of a surgical operation (mean for both new- and no-DNR patients: 20.8, p=0.978).
Table 1.
Baseline, clinical, and operative characteristics of matched new-DNR vs. no-DNR patients, ACS-NSQIP geriatric surgery pilot database (2014-2018)
| Characteristic | Matched Cohort |
||
|---|---|---|---|
| New-DNR (n=988) |
No-DNR (n=988) |
p value | |
| Age (years) | Matched | ||
| ≤75 | 271 (27.4%) | 275 (27.8%) | |
| 76 - 84 | 293 (29.7%) | 297 (30.1%) | |
| ≥ 85 | 424 (42.9%) | 416 (42.1%) | |
| Gender, female | 593 (60.0%) | 611 (61.8%) | 0.407 |
| Race | <0.001 | ||
| White | 639 (64.7%) | 805 (81.5%) | |
| Black | 75 (7.6%) | 86 (8.7%) | |
| Other/Unknown/Not Reported | 274 (27.7%) | 97 (9.8%) | |
| Functional status prior to surgery | <0.001 | ||
| Independent | 673 (68.1%) | 828 (83.8%) | |
| Partially dependent | 233 (23.6%) | 131 (13.3%) | |
| Totally dependent | 81 (8.2%) | 26 (2.6%) | |
| Unknown/Missing | 1 (0.1%) | 3 (0.3%) | |
| Transfer status | <0.001 | ||
| Admitted from home | 716 (72.5%) | 827 (83.7%) | |
| Acute care hospital | 65 (6.6%) | 33 (3.3%) | |
| Chronic care hospital | 132 (13.4%) | 65 (6.6%) | |
| Other | 75 (7.6%) | 63 (6.4%) | |
| History of a fall in the year prior to surgery | <0.001 | ||
| No | 359 (36.3%) | 439 (44.4%) | |
| Yes, within 3 months | 460 (46.6%) | 438 (44.3%) | |
| Yes, within 6 months | 55 (5.6%) | 19 (1.9%) | |
| Yes, within 1 year | 25 (2.5%) | 18 (1.8%) | |
| Unknown | 89 (9.0%) | 74 (7.5%) | |
| History of dementia, yes | 356 (36.0%) | 215 (21.8%) | <0.001 |
| Competency to sign surgical informed consent form | <0.001 | ||
| Patient signed own consent | 563 (57.0%) | 776 (78.5%) | |
| Consent signed by a surrogate | 395 (40.0%) | 195 (19.7%) | |
| Unknown | 30 (3.0%) | 17 (1.7%) | |
| Has advanced care plan prior to surgery, yes | 345 (34.9%) | 134 (13.6%) | <0.001 |
| Clinical and Operative Characteristics | |||
| Modified 5-Factor Frailty Index score (mFI-5) | Matched | ||
| 0 | 231 (23.4%) | 231 (23.4%) | |
| 0.2 | 584 (59.1%) | 584 (59.1%) | |
| 0.4 | 170 (17.2%) | 170 (17.2%) | |
| 0.6 | 3 (0.3%) | 3 (0.3%) | |
| Body Mass Index | 0.022 | ||
| Underweight | 85 (8.6%) | 69 (7.0%) | |
| Normal | 417 (42.2%) | 364 (36.8%) | |
| Overweight | 270 (27.4%) | 316 (32.0%) | |
| Obese | 201 (20.4%) | 215 (21.8%) | |
| Unknown | 15 (1.5%) | 24 (2.4%) | |
| ASA classification | <0.001 | ||
| 1-2 | 85 (8.6%) | 181(18.3%) | |
| 3-4 | 877 (88.8% | 803 (81.3%) | |
| 5 | 25 (2.5%) | 4 (0.4%) | |
| Specialty of Surgeon | 0.901 | ||
| Orthopedics | 474 (48.0%) | 467 (47.3%) | |
| General Surgery | 374 (37.9%) | 369 (37.3%) | |
| Vascular | 79 (8.0%) | 86 (8.7%) | |
| Other | 61 (6.2%) | 66 (6.7%) | |
| Surgery | <0.001 | ||
| Elective | 158 (16.0%) | 332 (33.6%) | |
| Urgent | 369 (37.3%) | 445 (45.0%) | |
| Emergent | 461 (46.7%) | 211 (21.4%) | |
Outcomes of new-DNR vs. no-DNR patients:
More than 75% of new-DNR patients experienced ≥1 postoperative complication as compared to 57.0% of matched no-DNR patients (p<0.001) (Table 2). For both new- and no-DNR patients, bleeding requiring transfusion(s) was the most common complication (Table 2), occurring within a median of 0 postoperative days for new- and 1 day for no-DNR patients (p=0.004). Cardiac arrest was uncommon, occurring in 4.9% of new- and 1.1% of no-DNR patients (p <0.001) (Table 2). Overall, new-DNR patients were more likely to undergo a reoperation within 30 days of their index surgery (8.6% vs. 3.6%, p<0.001).
Table 2.
Postoperative ≤30 day outcomes of matched new-DNR vs. no-DNR patients, ACS-NSQIP geriatric surgery pilot database (2014-2018)
| Outcome | Matched cohort |
||
|---|---|---|---|
| New-DNR (n=988) |
No-DNR (n=988) |
p value | |
| Any (≥1) complication 1 | 759 (76.8%) | 563 (57.0%) | <0.001 |
| Number of complication(s), median (IQR) | 2.0 (1.0, 3.0) | 1.0 (0.0, 1.0) | <0.001 |
| Most common complications | <0.001 | ||
| Bleeding requiring transfusions | 346 (35.0%) | 267 (27.0%) | |
| Prolonged ventilator use | 164 (16.6%) | 23 (2.3%) | |
| Pneumonia | 163 (16.5%) | 42 (4.25%) | |
| Septic shock | 151 (15.3%) | 22 (2.2%) | |
| Unplanned intubation | 142 (14.4%) | 21 (2.1%) | |
| Sepsis | 62 (6.3%) | 28 (2.8%) | |
| Myocardial infarction | 61 (6.2%) | 14 (1.4%) | |
| Cardiac arrest | 48 (4.9%) | 11 (1.1%) | |
| First complication | <0.001 | ||
| Bleeding requiring transfusions | 298 (30.2%) | 262 (26.5%) | |
| Septic shock | 77 (7.8%) | 15 (1.5%) | |
| Sepsis | 33 (3.3%) | 18 (1.8%) | |
| Pneumonia | 42 (4.3%) | 9 (0.9%) | |
| Unplanned intubation | 18 (1.8%) | 5 (0.5%) | |
| Prolonged ventilator use | 16 (1.6%) | 3 (0.3%) | |
| Cardiac arrest | 13 (1.3%) | 4 (0.4%) | |
| Other | 98 (9.9%) | 79 (25.1%) | |
| Postoperative delirium, yes | 425 (43.0%) | 248 (25.1%) | <0.001 |
| Pressure ulcer | <0.001 | ||
| No ulcer present at admission or discharge | 799 (80.9%) | 895 (90.6%) | |
| Progressed or new ulcer present at discharge | 127 (12.9%) | 30 (3.0%) | |
| Unknown | 62 (6.3%) | 63 (6.4%) | |
| Reoperation, yes | 85 (8.6%) | 36 (3.6%) | <0.001 |
| Readmission, yes | 80 (8.1%) | 112 (11.3%) | 0.015 |
| Postoperative hospice order, yes | 280 (28.3%) | 11 (1.1%) | <0.001 |
| Death | 388 (39.3%) | 38 (3.8%) | <0.001 |
| 30-day death or hospice (DoH) overall rate | 439 (44.4%) | 40 (4.0%) | <0.001 |
| Following Orthopedic surgery | 114 (24.1%) | 20 (4.3%) | <0.001 |
| Following General Surgery | 246 (65.8%) | 16 (4.3%) | <0.001 |
| Following Vascular surgery | 47 (59.5%) | 3 (3.5%) | <0.001 |
| Following Other surgical procedures | 32 (52.5%) | 1 (1.5%) | <0.001 |
| Following most common complication (bleeding) | 204 (59.1%) | 12 (4.5%) | <0.001 |
Occurence of any of the following complications: Superficial incisional surgical site infection; deep incisional surgical site infection; organ/space surgical site infection; wound disruption; pneumonia; unplanned intubation; pulmonary embolism; on ventilator greater than 48 hours; progressive renal insufficiency; acute renal failure; urinary tract infection; stroke/cerebrovascular accident; cardiac arrest; myocardial infarction; bleeding requiring transfusion(s); deep venous thrombosis requiring therapy; sepsis; septic shock; delirium; pressure ulcer development or progression
Patients with a new-DNR order experienced overall 30-day mortality that was approximately 10 times higher than that of matched no-DNR patients (39.3% vs 3.8%, p<0.001). Median days from surgery to death was 10 days for both new- and no-DNR patients (p=0.602). A postoperative hospice order was placed in 28.3% vs 1.1% of new- vs no-DNR patients (p<0.001), and was associated with a 30-day mortality rate of 81.8% for both patient groups (p = 0.988). Overall, the rate of DoH ≤30 postoperative days was 44.4% (new-DNR) vs. 4.0% (no-DNR) (p <0.001) but varied by surgical specialty (Table 2). The rate of DoH was similar for new-DNR patients undergoing elective vs. non-elective procedures (43.7% vs 44.5%, p= 0.833), but was significantly higher when compared to DoH rates for no-DNR patients undergoing elective (1.2%) vs non-elective procedures (5.5%), p=0.001.
Association between timing of DNR order and outcomes:
By phase of care, the rate of institution of new-DNR orders in the preoperative, day of surgery, and postoperative setting was 7.3%, 14.8%, and 55.2%., with a DoH rate of 16.7%, 23.3% and 64.6%, respectively (p<0.001). The association between a postoperative DNR order and high risk of DoH was observed across all specialties, with general surgery patients having higher rates of postoperative placement of DNR orders and DoH (Table 3). Location of the patient at the time of DNR order institution was also significant. The highest DoH rate was observed among patients with a new-DNR instituted in the ICU postoperatively (83.6%) (Figure 1).
Table 3.
Distribution of new-DNR patients and their rates of death or hospice transition (DoH) stratified by surgical specialty, ACS-NSQIP geriatric surgery pilot database (2014-2018)
| Specialty | Phase of care in which DNR order was placed |
Proportion of DNR patients* |
Rate of DoH |
|---|---|---|---|
| Orthopedic Surgery (n=474) | Preoperative | 13.0% | 17.7% |
| Day of surgery | 22.0% | 12.4% | |
| Postoperative | 37.6% | 41.0% | |
| Unknown | 27.2% | 13.2% | |
| General Surgery (n=374) | Preoperative | 2.7% | 10.0% |
| Day of surgery | 7.5% | 53.6% | |
| Postoperative | 74.3% | 77.0% | |
| Unknown | 15.5% | 27.6% | |
| Vascular (n=79) | Preoperative | - | - |
| Day of surgery | 12.7% | 60.0% | |
| Postoperative | 67.1% | 69.8% | |
| Unknown | 20.3% | 25.0% | |
| Other (n=61) | Preoperative | - | - |
| Day of surgery | 4.9% | 0.0% | |
| Postoperative | 62.3% | 76.3% | |
| Unknown | 32.8% | 15.0% | |
| Overall (n=988) | Preoperative | 7.3% | 16.7% |
| Day of surgery | 14.8% | 23.3% | |
| Postoperative | 55.2% | 64.5% | |
| Unknown | 22.7% | 17.9% |
Numbers may not add up to 100 due to rounding.
Figure 1.
Rate of death or hospice transition ≤ 30 postoperative days stratified by phase of care and patient location where the DNR order was instituted.
After adjusting for several risk factors in conditional multivariate analysis, a new-DNR order placed in any phase of care, as compared to a no-DNR order, was associated with a 28-fold likelihood of DoH (adjusted odds ratio: 28.1, 95% confidence interval: 13.0 - 60.1, p<0.001). In analysis including phase of care (Table 4), placement of a new DNR in postoperative period was associated with the highest odds of DoH. The three factors independently associated with DoH were (1) placement of a new-DNR order (particularly in the postoperative setting) (2) disseminated cancer, and (3) preoperative septic syndrome; the c-statistic of the model was 0.91. Of note, ASA class and urgency of surgery (elective vs. non-elective surgery) were not independently significant (Table 4) when a DNR order is included in the model. However, when the model is re-run without a DNR order, the risk factors that were independently associated with DoH were ASA class, urgency of surgery, sepsis, disseminated cancer, and cognitive competency at hospital admission.
Table 4.
Conditional multivariate logistic regression: Factors independently associated with postoperative death or hospice transition (DoH) ≤30 days in the matched study cohort, ACS-NSQIP geriatric surgery pilot database (2014-2018)
| Risk factor* | Adjusted Odds Ratio (95% Confidence Interval) |
p-value |
|---|---|---|
| New-DNR order | ||
| No order | Reference | |
| Order placed preoperatively or on the day of surgery | 5.8 (1.8, 18.9) | 0.003 |
| Order placed postoperatively | 52.9 (20.0, 139.8) | <0.001 |
| Order placed in unknown phase of care | 7.7 (2.0, 29.7) | 0.003 |
| Female vs male | 0.5 (0.3, 1.2) | 0.125 |
| Race | ||
| White | Reference | |
| Black | 2.9 (0.6, 14.2) | 0.186 |
| Other/Unknown | 1.0 (0.3, 3.4) | 0.985 |
| Transfer status | ||
| Admitted from home | Reference | |
| Acute care hospital | 0.8 (0.2, 4.2) | 0.817 |
| Chronic care hospital | 1.5 (0.5, 4.4) | 0.508 |
| Other | 1.1 (0.2, 5.3) | 0.884 |
| Cognitive competency on admission | ||
| Patient signed own consent | Reference | |
| Consent signed by a surrogate | 1.8 (0.7, 4.8) | 0.212 |
| Unknown | 2.1 (0.2, 18.4) | 0.503 |
| BMI | ||
| Underweight | 3.8 (0.9, 16.5) | 0.074 |
| Normal | Reference | |
| Overweight | 0.6 (0.2, 1.7) | 0.324 |
| Obese | 2.1 (0.6, 6.7) | 0.221 |
| Unknown | 0.1 (0.0, 1.1) | 0.056 |
| Disseminated cancer | 31.0 (5.6, 172.8) | <0.001 |
| Systemic sepsis/shock | 5.5 (1.5, 19.7) | 0.009 |
| Coagulopathy | 0.7 (0.2, 1.8) | 0.419 |
| Urgency of surgery | ||
| Elective | Reference | |
| Urgent | 2.9 (0.6, 13.5) | 0.182 |
| Emergent | 3.0 (0.5, 18.0) | 0.240 |
| ASA class: 4-5 vs 1-3 | 2.2 (0.9, 5.6) | 0.104 |
Because data were matched by age, mFI-5 (modified frailty index score), and procedure, these factors were not separately included in the conditional regression model.
DISCUSSION
Using a large multi-institutional dataset that includes geriatric care-specific data not commonly found in other databases, we examined the characteristics and outcomes of older adults who had an in-hospital code status change to DNR during their surgical admission. In the initial sample of more than 41000 patients, 2.8% of older adults had a new-DNR order, but this increased with age to approximately 10% among patients aged ≥85 years. In the matched sample of new-DNR and no-DNR patients, preoperative frailty scores were often low with majority of patients having scores ≤0.2. In adjusted analysis, patients with a new-DNR order were 28 times more likely to experience DoH ≤30 postoperative days than those without a DNR order. When considering phase of care, the highest risk was among those whose DNR order was placed postoperatively. Factors frequently associated with poor surgical outcomes such as ASA class and non-elective surgery were not independently associated with DoH. Our results indicate that (1) older patients who have an in-hospital code status change to DNR are a subset of extremely high-risk patients and (2) traditional risk assessment measures alone may not effectively identify those who may decompensate critically and transition to DoH in the early postoperative period. This study is consistent with other research documenting association of DNR order with poor outcomes [13, 14, 18, 20, 21, 23, 24] and extends previous work by identifying variation in postoperative outcomes based on timing of DNR placement. Overall, this study highlights potential gaps in perioperative prognostication, which may have significant implications on goals of care discussions, surgical decision- making, and patient outcomes.
From a risk-stratification perspective, and to the extent that a DNR order serves as a marker of risk, this study revealed two distinct subsets of new-DNR patients. The first subset includes those who had a DNR order placed “early” i.e. preoperatively or on the day of surgery. With the exception of patients who had this order instituted in the ICU on the day of surgery (indicating critical illness), the mortality rate of other patients in this subset was within the 12.5% - 36.9% mortality rate for patients with pre-admission DNR orders undergoing diverse procedures published in the literature [13-14, 18, 20-21]. In contrast, the larger subset of patients who had a new-DNR order placed postoperatively had a 64.5% DoH rate, with dramatic odds of DoH which was more than 50-fold higher than for those without a DNR order, and approximately 10-fold higher than for those with a preoperative new-DNR order. For this subset of DNR patients, their high-risk status may have become evident in the postoperative setting following unexpected and significant clinical deterioration.
Whereas clinical decline may signify limited physiologic reserve of older adults, variations in outcomes associated with the presence of DNR order and the phase of care when placed suggest that DNR orders are a marker of risk beyond traditional prognostic measures. A number of findings support this conclusion: firstly, with the inclusion of frailty score in our matching algorithm, we expected relatively comparable outcomes between new-and no-DNR patients. Instead, we found substantial differences in 30-day outcomes. Secondly, since urgent and emergent surgery (as compared to elective surgery) are traditionally associated with higher complication and mortality rates, we expected worse outcomes following non-elective cases; we found no significant difference in rate of DoH transition for new-DNR patients undergoing elective vs. non-elective procedures. Moreover, multivariate analysis showed that neither surgical urgency (elective or non-elective surgery) nor preoperative ASA class were independently associated with DoH. Overall, these findings suggest that the very act of surgical intervention may be a critical tipping point for some high-risk older adults, and that traditional markers of poor postoperative outcomes may not effectively identify those at risk of DoH postoperatively.
This study has a number of notable limitations. Details of patient-provider discussions of goals and transitions of care are not collected in ACS-NSQIP, therefore it is impossible to determine key factors that influenced the placement of a new-DNR order. Details on provider and hospital characteristics such as teaching status are not included in ACS-NSQIP, and our results may not be generalizable given the relatively small number of hospitals that provided data for the ACS-NSQIP geriatric database. The database does not include granular information on medications, intraoperative events, severity of illness and comorbidities, as well severity of complications (such as units of blood products transfused into patients experiencing bleeding). Although some studies have indicated an association between frailty and peri-procedural bleeding risk [30-31], it is unclear why a sizeable proportion of older adults in this study experienced perioperative bleeding requiring transfusions, and why new-DNR patients experienced worse outcomes following this complication. Nonetheless, our results underscore the critical importance of assessing coagulopathy history and therapies, intraoperative hemostasis, and close monitoring of coagulation status postoperatively in geriatric surgical care. Variations in intensity of postoperative care or use of other life-sustaining therapies based on presence and timing of DNR may explain improved outcomes of patients who had a preoperative DNR order as compared to those who had the order placed postoperatively; this cannot be ascertained in the database. A multidisciplinary approach to surgical care has been shown to improve outcomes of older adults undergoing surgery [32-33], and may explain why patients undergoing orthopedic surgery- a specialty that often utilizes a co-management model – experienced lower rates of DoH. Whereas the process leading to discussions and placement of a DNR order may vary by surgical service, this information, as well as the use of specific care pathways, and participation of patients in prehabilitation programs are not included in the ACS-NSQIP database. Cost of care, outcomes >30days, and geriatric-specific measures such as gait speed, grip strength, and failure to thrive are also not captured in ACS-NSQIP. For patients who had a DNR order instituted on the day of surgery, it is impossible to determine whether the order was placed before or after surgery; details on rescindment and reinstatement of a DNR order are also unavailable. In addition, we cannot ascertain whether high rates of DoH for new-DNR patients represent a lower likelihood to consent to aggressive interventions. Interestingly, new-DNR patients in this study had higher rates of reoperation, implying consent for aggressive or additional interventions in the initial stages of their care.
In conclusion, our study demonstrates that older adults who have an inpatient DNR order placed during their surgical care often experience poor surgical outcomes, and that a DNR order is associated with death or transition to hospice even when traditional markers of risk are controlled. Further work is needed to delineate the unmeasured factors associated with DNR orders and aid shared decision-making based on goals of care. Overall, more studies are needed to fine-tune perioperative risk assessment of older adults to enable early identification of extremely high-risk patients, who may not be great candidates for resource-intense operative interventions.
Supplementary Material
Supplemental Table S1: Variables* analyzed in this study that are exclusively in the ACS-NSQIP Geriatric Surgery pilot database (2014-2018)
*Definitions of all other ACS-NSQIP variables (demographic, clinical, operative characteristics, and complications) analyzed in this study are available at https://www.facs.org/-/media/files/quality-programs/nsqip/nsqip_puf_userguide_2018.ashx
Supplemental Table S2: List of the 18 complications (excluding death, reoperation, and readmission) captured in the ACS-NSQIP participant use file.
Supplemental Table S3: Additional clinical characteristics of matched new-DNR vs. no-DNR patients, ACS-NSQIP geriatric surgery pilot database (2014-2018)
Key Points
An inpatient DNR order is associated with high rates of 30-day death and hospice transition, especially in the postoperative setting.
Why Does This Paper Matter?
Better perioperative risk assessment of older adults is needed to enable early identification of extremely high-risk patients.
ACKNOWLEDGMENT:
The Duke Biostatistics, Epidemiology, and Research Design Methods Core’s support of this project was made possible in part by CTSA Grant (UL1TR002553) from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH), and the NIH Roadmap for Medical Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NCATS or NIH.
Funding:
This study was supported by a grant from the Duke Center for Geriatric Surgery awarded to Hadiza S Kazaure, MD.
Sponsor’s Role:
There were no sponsors for this project.
Footnotes
Conflict(s) of Interest: The authors have no conflict of interest.
Disclosures: The American College of Surgeons National Surgical Quality Improvement Program and the hospitals participating in the ACS NSQIP are the source of the data used herein; they have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.
REFERENCES
- 1.McDermott KW (IBM Watson Health), Freeman WJ (AHRQ), Elixhauser A (AHRQ). Overview of Operating Room Procedures During Inpatient Stays in U.S. Hospitals, 2014. HCUP Statistical Brief #233. December 2017. Agency for Healthcare Research and Quality, Rockville, MD. www.hcup-us.ahrq.gov/reports/statbriefs/sb233-Operating-Room-Procedures-United-States-2014.pdf. Accessed September 18, 2020. [PubMed] [Google Scholar]
- 2.Dall TM, Gallo PD, Chakrabarti R, et al. An aging population and growing disease burden will require a large and specialized health care workforce by 2025.Health Aff 2013; 32: pp. 2013–2020. [DOI] [PubMed] [Google Scholar]
- 3.Turrentine FE,Wang H, Simpson VB, et al. Surgical risk factors, morbidity, and mortality in elderly patients. J Am Coll Surg 2006;203:865–77. [DOI] [PubMed] [Google Scholar]
- 4.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. [DOI] [PubMed] [Google Scholar]
- 5.Hall DE, Arya S, Schmid KK, et al. Association of a frailty screening initiative with postoperative survival at 30, 180, and 365 days. JAMA Surg 2017; 152: pp. 233–240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Joseph B, Zangbar B, Pandit V, et al. Emergency general surgery in the elderly: too old or too frail?.J Am Coll Surg 2016; 222: pp. 805–813 [DOI] [PubMed] [Google Scholar]
- 7.Saxton A, Velanovich V. Preoperative frailty and quality of life as predictors of postoperative complications. Ann Surg, 253 (2011), pp. 1223–1229 [DOI] [PubMed] [Google Scholar]
- 8.Robinson TN, Walston JD, Brummel NE, et al. Frailty for surgeons: Review of a National Institute on Aging Conference on Frailty for Specialists. J Am Coll Surg, 221 (2015), pp. 1083–1092 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berian JR, Rosenthal RA. Redefining quality of surgical care for the frail elderly. Curr Surg Rep. 2016;4:1–11. [Google Scholar]
- 10.Hewitt J, Moug SJ, Middleton M, Chakrabarti M, Stechman MJ, McCarthy K. Prevalence of frailty and its association with mortality in general surgery. Am J Surg. 2015;209(2):254–259. doi: 10.1016/j.amjsurg.2014.05.022. [DOI] [PubMed] [Google Scholar]
- 11.Lin H-S, Watts JN, Peel NM, Hubbard RE. Frailty and post-operative outcomes in older surgical patients: a systematic review. BMC Geriatr. 2016;16(1):157. doi: 10.1186/s12877-016-0329-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kim S, Han H, Jung H, et al. Multidimensional frailty score for the prediction of postoperative mortality risk. JAMA surgery. 2014. Jul;149(7):633–40 [DOI] [PubMed] [Google Scholar]
- 13.Kazaure HS, Roman SA, Sosa JA. High mortality in surgical patients with Do-Not-Resuscitate (DNR) Orders: Analysis of 8,256 patients. Arch Surg. 2011. Aug; 146(8):922–8. [DOI] [PubMed] [Google Scholar]
- 14.Saager L, Kurz A, Deogaonkar A, et al. Pre-existing do-not-resuscitate orders are not associated with increased postoperative morbidity at 30 days in surgical patients. Crit Care Med 2011;39:1036–41. [DOI] [PubMed] [Google Scholar]
- 15.Kazaure HS, Roman SA, Rosenthal RA, Sosa JA. Cardiac arrest among surgical patients: an analysis of incidence, patient characteristics, and outcomes in ACS-NSQIP. JAMA Surg. 2013;148(1):14–21 [DOI] [PubMed] [Google Scholar]
- 16.Walkey AJ, Weinberg J, Wiener RS, Cooke CR, Lindenauer PK. Association of Do-Not-Resuscitate Orders and Hospital Mortality Rate Among Patients With Pneumonia. JAMA Intern Med. 2016. Jan;176(1):97–104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Yang AD, Bentrem DJ, Pappas SG, et al. Advance directive use among patients undergoing high-risk operations. Am. J. Surg. Jul 2004;188(1):98–101. [DOI] [PubMed] [Google Scholar]
- 18.Speicher PJ, Lagoo-Deenadayalan SA, Galanos AN, et al. Expectations and outcomes in geriatric patients with do-not-resuscitate orders undergoing emergency surgical management of bowel obstruction. JAMA Surg, 148 (2013), pp. 23–28 [DOI] [PubMed] [Google Scholar]
- 19.Meissner A, Genga KR, Studart FS, et al. Epidemiology of and factors associated with end-of-life decisions in a surgical intensive care unit. Crit Care Med 2010;38:1060–8. [DOI] [PubMed] [Google Scholar]
- 20.Scarborough JE, Pappas TN. The effect of do-not-resuscitate status on postoperative mortality in the elderly following emergency surgery. Adv Surg 2013; 47: 213–25 [DOI] [PubMed] [Google Scholar]
- 21.Brovman EY, Walsh EC, Burton BN, et al. Postoperative outcomes in patients with a do-not-resuscitate (DNR) order undergoing elective procedures. J Clin Anesth. 2018;48:81–88. doi: 10.1016/j.jclinane.2018.05.007 [DOI] [PubMed] [Google Scholar]
- 22.Beverly A, Brovman EY, Urman RD. Comparison of Postoperative Outcomes in Elderly Patients With a Do-Not-Resuscitate Order Undergoing Elective and Nonelective Hip Surgery. Geriatr Orthop Surg Rehabil. 2017;8(2):78–86. doi: 10.1177/2151458516685826 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Siracuse JJ, Jones DW, Meltzer EC et al. Impact of “Do Not Resuscitate” status on the outcome of major vascular surgical procedures. Ann Vasc Surg 2015; 29: 1339–45 [DOI] [PubMed] [Google Scholar]
- 24.Maxwell BG, Lobato RL, Cason MB, Wong JK. Perioperative morbidity and mortality of cardiothoracic surgery in patients with a do-not-resuscitate order. PeerJ. 2014;2:e245. Published 2014 Jan 22. doi: 10.7717/peerj.245 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Subramaniam S, Aalberg JJ, Soriano RP, Divino CM. A new five factor modified frailty index using NSQIP data. J Am Coll Surg, 226 (2018), pp. 173–181 [DOI] [PubMed] [Google Scholar]
- 26.Tracy Brett M., Adams Mata A., Schenker Mara L., Gelbard Rondi B.. The 5 and 11 Factor Modified Frailty Indices are Equally Effective at Outcome Prediction Using TQIP, J Surg Research. 2020. 255, pp 456–462. 10.1016/j.jss.2020.05.090. [DOI] [PubMed] [Google Scholar]
- 27.Ali TZ, Lehman EB, Aziz F. Modified frailty index can be used to predict adverse outcomes and mortality after lower extremity bypass surgery. Ann Vasc Surg, 46 (2018), pp. 168–177 [DOI] [PubMed] [Google Scholar]
- 28.Traven SA, Horn RW, Reeves RA, Walton ZJ et al. The 5-factor modified frailty index predicts complications, hospital admission, and mortality following arthroscopic rotator cuff repair. Arthrosc J Arthrosc Relat Surg. 2020. 36 pp. 383–388 [DOI] [PubMed] [Google Scholar]
- 29.American College of Surgeons National Surgical Quality Improvement Program. ACS-NSQIP user guide for the 2014-18 participant data use files. https://www.facs.org/quality-programs/acs-nsqip/participant-use Accessed December 15, 2020
- 30.Dodson JA, Hochman JS, Roe MT, et al. The Association of Frailty With In-Hospital Bleeding Among Older Adults With Acute Myocardial Infarction: Insights From the ACTION Registry [published correction appears in JACC Cardiovasc Interv. 2018;11(22):2287–2296. doi: 10.1016/j.jcin.2018.08.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Tang M, Yu C, Hu P, Wang C, Sheng J, Ma S. Risk factors for bleeding after dental extractions in patients over 60 years of age who are taking antiplatelet drugs. Br J Oral Maxillofac Surg. 2018;56(9):854–858. doi: 10.1016/j.bjoms.2018.09.009 [DOI] [PubMed] [Google Scholar]
- 32.McDonald SR, Heflin MT, Whitson HE, Dalton TO, Lidsky ME, Liu P, Poer CM, Sloane R, Thacker JK, White HK, Yanamadala M, Lagoo-Deenadayalan SA. Association of Integrated Care Coordination With Postsurgical Outcomes in High-Risk Older Adults: The Perioperative Optimization of Senior Health (POSH) Initiative. JAMA Surg. 2018. May 1;153(5):454–462. doi: 10.1001/jamasurg.2017.5513. PMID: 29299599; PMCID: PMC5875304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Grigoryan KV, Javedan H, Rudolph JL. Orthogeriatric care models and outcomes in hip fracture patients: a systematic review and meta-analysis. J Orthop Trauma. 2014. Mar;28(3):e49–55. doi: 10.1097/BOT.0b013e3182a5a045. PMID: 23912859; PMCID: PMC3909556. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Table S1: Variables* analyzed in this study that are exclusively in the ACS-NSQIP Geriatric Surgery pilot database (2014-2018)
*Definitions of all other ACS-NSQIP variables (demographic, clinical, operative characteristics, and complications) analyzed in this study are available at https://www.facs.org/-/media/files/quality-programs/nsqip/nsqip_puf_userguide_2018.ashx
Supplemental Table S2: List of the 18 complications (excluding death, reoperation, and readmission) captured in the ACS-NSQIP participant use file.
Supplemental Table S3: Additional clinical characteristics of matched new-DNR vs. no-DNR patients, ACS-NSQIP geriatric surgery pilot database (2014-2018)