Abstract
Background:
Pursuing pancreatic resection in elderly patients is often complex and limited by concern for functional status and postoperative risk. This study examines the associations between two different preoperative functional status metrics with postoperative outcomes in the geriatric population.
Methods:
Patients who participated in the ACS NSQIP Geriatric Surgery Research File pilot program (2014–2018) undergoing elective pancreatic operations were included. Two clinically meaningful functional status scores were calculated: presence of one or more geriatric-specific variable (GSV) and a 5-factor modified frailty index (mFI-5). Multivariable logistic regression adjusting for ACS NSQIP estimated risk was performed to evaluate associations between preoperative GSV, mFI-5 and 30-day outcome measures.
Results:
1266 patients were included: 808 (64%) age 65-74, 302 (24%) age 75-80, and 156 (12%) age ≥ 81; 843 (67%) patients underwent pancreatoduodenectomy. Operations were performed for pancreatic adenocarcinoma in 712 (56%) patients. Older patients had greater likelihood of postoperative morbidity (35% vs 31% vs 47%, by age group, p = 0.004) and discharge to a facility (12% vs 23% vs 48%, by age group, p < 0.001). Adjusting for ACS NSQIP predicted risk, patients with a preoperative GSV were more likely to require reoperation and discharge to a facility (OR 1.81 [95% CI 1.03-3.16] and 3.95 [95% CI 2.91-5.38], respectively). The mFI-5 was not associated with postoperative outcomes (all p ≥ 0.18).
Conclusion:
Presence of a preoperative GSV is associated with reoperation and discharge to a skilled facility following elective pancreatic resection. Geriatric-specific variables should be considered in joint preoperative decision making to optimize care.
Introduction:
Patients 65 years and older represent the fastest expanding population group in the United States.1 The prevalence of pancreatic cancer and premalignant pancreatic neoplasms increases with age; postoperative complications and mortality similarly increase with age.2–5 As such, preoperative joint decision making focused on weighing patients’ preoperative health state with postoperative risk are critically important as we continue to strive to improve the quality of care for this patient population.6–8
A number of various patient factors, such as age, sarcopenia, and frailty, have been associated with postoperative morbidity and mortality in geriatric patients.9–12 The comprehensive Optimal Perioperative Management of the Geriatric Patient created by the American College of Surgeons (ACS) was developed to aid with risk stratification in this population, however the complexity and number of items in this tool can be burdensome to the user which has limited its clinical applicability.13 The 5-Factor Modified Frailty Index (mFI-5) is a more recent attempt to quantify frailty and is composed of five National Surgical Quality Improvement Program (NSQIP) reported variables – functional status, diabetes, history of chronic obstructive pulmonary disease, history of congestive heart failure, and hypertension requiring medication. The mFI-5 was shown to have predictive capabilities for postoperative complications and mortality in general surgery and in the geriatric population14,15, but not among pancreatic surgery patients. A potential weakness of the mFI-5 is its focus on comorbid conditions and lack of inclusion of variables that seek to describe the frail state.16 While comorbid conditions should be strongly considered in risk stratification for elderly patients, the presence of frailty can have an enormous impact on the surgical risk of this population.17,18
The ACS NSQIP Geriatric Surgery Pilot Program was launched in 2014 and one of the goals was to obtain data that reflects the needs of the geriatric population. This pilot program was a collaboration between the ACS NSQIP and ACS Geriatric Task Force. Data has been collected since 2014 and includes 33 hospitals with over 60,000 patients. A number of geriatric-specific variables (GSV) were collected that addressed traits associated with a frail state—these variables fall into categories including cognition, capacity for decision making, functional dependence, and mobility.19 The primary aim of this study was to investigate the association between GSVs and postoperative morbidity and mortality in geriatric patients selected for elective pancreatic operations. The secondary aim was to evaluate associations between preoperative GSVs and postoperative reoperation, discharge disposition, and readmissions. In addition, we tested associations between mFI-5 and the same postoperative outcomes in this patient population to compare performance of GSVs versus mFI-5 in preoperative risk assessment for geriatric pancreatic surgery patients.
Methods:
Data and Study Population:
All patients ≥ 65 years of age who underwent elective pancreatic operations between January 1, 2014 and December 31, 2018 and captured in the ACS NSQIP Geriatric Surgery Research File (GSRF) were included in the study. Cases from the GSRF were linked using patients’ unique case ID with the ACS NSQIP Participant Use Data File (PUF) and Pancreatectomy Procedure Targeted PUF. The ACS NSQIP PUF, Pancreatectomy Procedure Targeted PUF, and GSRF report Health Insurance Portability and Accountability Act of 1966 (HIPAA) compliant aggregate patient data from participating institutions. Patients < 65 years of age and patients who underwent emergent operations were excluded. Analysis of these ACS NSQIP de-identified datasets was reviewed and designated exempt by the University of Virginia Review Board for Health Sciences Research.
Relevant demographic variables from the data sets included age, sex, and race/ethnicity. Clinical covariates included comorbidities, American Society of Anesthesiologists (ASA) classification, functional status (categorized as independent or dependent), Body Mass Index (BMI) (kg/m2), diagnosis (categorized as pancreatic adenocarcinoma, chronic pancreatitis, or other), albumin, operation (defined by Current Procedural Terminology [CPT] codes: 48140, 48145, 48146, 48148, 48150, 48152, 48153, 48154, 48155, 48999), and ACS NSQIP estimated probability of morbidity and mortality. Clinical GSVs included: patient origin status (living home alone, living at home with family and/or friends, or living at a facility other than home), use of a mobility aid, fall history (categorized as no falls or at least one fall within the past year), and competency on admission (categorized as able to sign surgical consent form or unable to sign consent). Patient origin status was categorized as high risk (lives home alone or lives at a facility other than home) or low risk (lives at home with family and/or friends). GSVs were modeled as a categorical variable (≥ 1 present vs none present). Modified 5-Item Frailty Index (mFI-5) scores were calculated for each patient as described previously based on defined comorbid conditions.14
Primary outcome measures were defined as ACS NSQIP defined composite morbidity and mortality. ACS NSQIP outcomes are summarized as events occurring during the index hospitalization or within the 30 days after the operation. ACS NSQIP defined composite morbidity includes pneumonia; reintubation; failure to wean off the ventilator within 48 hours; renal insufficiency; renal failure; cardiac arrest; myocardial infarction; cerebral vascular accident; sepsis; septic shock; fascial dehiscence; superficial, deep, or organ–space surgical site infection; urinary tract infection; bleeding requiring perioperative blood transfusion; venous thromboembolism (including deep venous thrombosis and/or pulmonary embolus). Postoperative pancreatic fistula was added to the composite morbidity variable.
Secondary outcome measures were defined as the need for reoperation within 30 days of the index operation, discharge destination, and unplanned readmission within 30 days of the index operation. Discharge destination was categorized as either home or discharge to a facility other than home (e.g. separate acute care facility, skilled care, unskilled care facility, rehabilitation facility, and other non-home destinations).
Statistical Analysis:
Continuous data is presented as a median (interquartile range [IQR]); categorical variables are reported as frequency (percent); age is reported as categorical variable groups (65 – 74, 75 – 80, and ≥81). Univariate comparisons were performed using Chi-squared, Fisher’s exact test, Student’s t-test, or Kruskal-Wallis test, as appropriate.
Logistic regression models were created with either GSVs or the mFI-5 score as the predictor variables of the outcome measures. The models were adjusted for ACS NSQIP-estimated probability of mortality (for the mortality model) or the ACS NSQIP-estimated probability of morbidity (for the morbidity, reoperation, discharge destination, and readmission models) to adjust for baseline patient-specific risk associated with their individual comorbid condition profile. Additional logistic regression models tested associations between individual GSV and each of the primary and secondary outcome measures. Patients whose origin status was not from home were excluded from the models testing associations with discharge destination.
Proportions of missing data among variables were low. Fall history variable had the greatest proportion of missing data (8%). When present, proportion of missing data in remaining variables ranged from 0.2% to 2%. Patients with missing data were removed from analyses that utilized the missing variable. All hypothesis tests were two-tailed with an alpha value of 0.05. Data management and statistical analyses were performed using SAS v 9.4 (SAS Institute, Cary, NC).
Results:
Patient characteristics
A total of 1266 patients were included in the study: 808 (64%) age 65-74, 302 (24%) age 75-80, 156 (12%) age ≥ 81. A total of 843 (67%) patients underwent pancreatoduodenectomy. Compared to patients aged 65-74 and 75-80, patients ≥ 81 years had a higher ACS NSQIP-calculated probability of morbidity (25.3% [18.4–30.2], 26.0% [19.3–32.7], 29.8% [22.2–37.3], p<0.001) and mortality (1.1% [0.6–2.0], 1.7% [0.8–3.1], 2.7% [1.6–5.7], p < 0.001). Demographics and clinical covariates stratified by age group are summarized in Table 1. The median mFI-5 score was not different between age groups (p = 0.82). When compared to patients aged 65-74, patients in the 75-80, and ≥ 81 age groups were more likely to have at least one GSV present prior to resection (29% vs 49% vs 51%, p < 0.001), though the average number of GSVs present did not differ between age groups (1.00 vs 1.03 vs 1.05, p = 0.56). A total of 806 (63.7%) had no GSVs, 364 (28.8%) had 1 GSV, 83 (6.6%) had 2 GSVs, 12 (1.0%) had 3 GSVs, and 1 (0.1%) had 4 GSVs. Individual variable components of GSV and mFI-5 composite scores are summarized in Table 2.
Table 1.
Demographic and clinical variables by age group
| Age 65–74 (n = 808) |
Age 75–80 (n = 302) |
Age ≥ 81 (n = 156) |
p-value | |
|---|---|---|---|---|
| Male sex | 402 (49.8) | 136 (45.0) | 95 (60.9) | 0.006 |
| Race / ethnicity | 0.05 | |||
| White | 679 (84.0) | 248 (82.1) | 135 (86.5) | |
| Black | 78 (9.7) | 26 (8.6) | 12 (7.7) | |
| Hispanic | 17 (2.1) | 6 (2.0) | 6 (3.9) | |
| Asian / American Indian | 11 (1.4) | 13 (4.3) | 2 (1.3) | |
| Unknown / not reported | 23 (2.9) | 9 (3.0) | 1 (0.6) | |
| Smoking | 133 (10.5) | 103 (12.7) | 22 (7.3) | 0.002 |
| BMI, mean (SD) | 27.4 (6.0) | 26.9 (4.9) | 25.6 (4.5) | < 0.001 |
| ASA Class | 0.005 | |||
| 1 | 1 (0.1) | 0 (0) | 0 (0) | |
| 2 | 153 (19.0) | 50 (16.6) | 15 (9.6) | |
| 3 | 609 (75.5) | 223 (73.8) | 128 (82.1) | |
| 4 | 44 (5.4) | 29 (9.6) | 13 (8.3) | |
| Albumin | 3.9 | 3.8 | 3.6 | < 0.001 |
| Diagnosis | 0.002 | |||
| Pancreatic adenocarcinoma* | 432 (53.5) | 170 (56.3) | 110 (70.5) | |
| Chronic pancreatitis | 33 (4.1) | 15 (5.0) | 0 (0) | |
| Other† | 343 (42.4) | 117 (38.7) | 46 (29.5) | |
| Pancreaticoduodenectomy | 538 (66.6) | 188 (62.3) | 117 (75.0) | 0.02 |
| ACS NSQIP estimated probability of morbidity, % | 25.3 (18.4–30.2) | 26.0 (19.3–32.7) | 29.8 (22.2–37.3) | < 0.001 |
| ACS NSQIP estimated probability of mortality, % | 1.1 (0.6–2.0) | 1.7 (0.8–3.1) | 2.7 (1.6–5.7) | < 0.001 |
BMI Body mass index, ASA American Society of Anesthesiologists, ACS NSQIP American College of Surgeons National Surgical Quality Improvement Program
Pancreatic adenocarcinoma includes pancreatic ductal adenocarcinoma and invasive intraductal papillary mucinous neoplasm (IPMN)
Other: ampullary carcinoma, cystadenocarcinoma, distal cholangiocarcinoma, duodenal carcinoma, mucinous cystic neoplasm, non-invasive IPMN, neuroendocrine neoplasms, mucinous, serous cystadenoma, pathology not originating from the pancreas, unknown histology
Table 2.
Geriatric-specific variable and 5-factor modified frailty index components
| Age 65–74 (n = 808) |
Age 75–80 (n = 302) |
Age ≥ 81 (n = 156) |
p-value | |
|---|---|---|---|---|
| GSV | 232 (28.7) | 148 (49.0) | 80 (51.3) | < 0.001 |
| Lives home alone or at a facility | 135 (16.9) | 85 (28.4) | 49 (32.0) | < 0.001 |
| Uses mobility aid | 86 (10.8) | 60 (20.1) | 40 (26.3) | < 0.001 |
| Fall history | 50 (6.8) | 33 (11.7) | 17 (11.7) | 0.02 |
| Did not sign own consent | 7 (0.9) | 5 (1.7) | 3 (2.0) | 0.36 |
| mFI-5 score, median (IQR) | 0.20 (0–0.40) | 0.20 (0 – 0.40) | 0.20 (0.20–0.20) | 0.82 |
| History of CHF | 3 (0.4) | 1 (0.3) | 0 (0) | 0.75 |
| Diabetes | 231 (28.6) | 76 (25.2) | 37 (23.7) | 0.30 |
| History of COPD | 44 (5.4) | 19 (6.3) | 8 (5.1) | 0.83 |
| Hypertension requiring medication | 522 (64.6) | 211 (69.9) | 108 (69.2) | 0.19 |
| Dependent functional status | 6 (0.7) | 6 (2.0) | 4 (2.6) | 0.07 |
GSV geriatric-specific variable, mFI-5 5-factor modified frailty index, IQR interquartile range, CHF congestive heart failure, COPD chronic obstructive pulmonary disease
Outcome measures
Postoperative primary and secondary outcome measures are summarized in Table 3. A total of 456 (36%) patients had a postoperative composite morbidity. There was no difference in postoperative morbidity between age strata (41% vs 39% vs 50%, by age group, p = 0.06). There was no difference in postoperative proportions of pancreatic fistulas (16.9% vs 14.6% vs 13.5%, by age group, p = 0.64) or length of stay (8.6 vs 10.2 vs 7.8 days, by age group, p = 0.13). A total of 29 (2.3%) patients died in the 30-day postoperative period: 14 (1.7%) 65-74 years, 10 (3.3%) 75-80 years, 5 (3.2%) ≥ 81 years; there was no difference in mortality by age group, p = 0.21. Patients age 75-80 and ≥81 years were significantly more likely to be discharged to a facility (12% vs 23% vs 48%, by age group, p < 0.001). There was no difference in the proportion of reoperations or 30-day readmissions between patients in different age categories (both p≥0.34).
Table 3.
Postoperative outcomes by age group among the entire patient cohort
| Age 65–74 (n = 808) |
Age 75–80 (n = 302) |
Age ≥ 81 (n = 156) |
p-value | |
|---|---|---|---|---|
| Composite morbidity | 337 (41.7) | 120 (39.7) | 79 (50.6) | 0.06 |
| Mortality | 14 (1.7) | 10 (3.3) | 5 (3.2) | 0.21 |
| Return to the operating room | 35 (4.3) | 13 (4.3) | 4 (2.6) | 0.58 |
| Discharge to facility | 91 (11.5) | 67 (22.5) | 73 (47.7) | < 0.001 |
| Unplanned readmission | 138 (17.1) | 42 (13.9) | 29 (18.6) | 0.34 |
| Pneumonia | 25 (3.1) | 12 (4.0) | 11 (7.0) | 0.059 |
| Reintubation | 29 (3.6) | 12 (4.0) | 12 (7.7) | 0.063 |
| Failure to wean ventilator | 19 (2.3) | 7 (2.3) | 6 (3.8) | 0.53 |
| MI | 10 (1.2) | 2 (0.7) | 3 (1.9) | 0.48 |
| Cardiac arrest | 10 (1.2) | 4 (1.3) | 1 (0.6) | 0.79 |
| Transfusion | 121 (15.0) | 42 (13.9) | 33 (21.1) | 0.10 |
| DVT | 18 (2.2) | 2 (0.7) | 3 (1.9) | 0.22 |
| PE | 8 (1.0) | 1 (0.3) | 0 (0) | 0.27 |
| Sepsis | 33 (4.1) | 15 (5.0) | 12 (7.7) | 0.15 |
| Septic shock | 22 (2.7) | 13 (4.3) | 6 (3.8) | 0.37 |
| SSSI | 57 (7.0) | 16 (5.3) | 11 (7.0) | 0.56 |
| DSSI | 6 (0.7) | 2 (0.7) | 1 (0.6) | 0.98 |
| OSSI | 96 (11.9) | 27 (7.9) | 15 (9.6) | 0.15 |
| Fascial dehiscence | 6 (0.7) | 2 (0.7) | 3 (1.9) | 0.31 |
| Renal insufficiency | 2 (0.2) | 1 (0.3) | 1 (0.6) | 0.72 |
| Acute renal failure | 6 (0.7) | 3 (1.0) | 1 (0.6) | 0.89 |
| Urinary tract infection | 19 (2.3) | 11 (3.6) | 8 (5.1) | 0.13 |
| Stroke | 1 (0.1) | 1 (0.3) | 1 (0.6) | 0.44 |
| Pancreatic fistula | 130 (16.9) | 44 (14.6) | 21 (13.5) | 0.64 |
| Length of stay (days), mean (SD) | 9.23 (8.6) | 9.69 (10.2) | 10.79 (7.8) | 0.13 |
MI myocardial infarction, DVT deep venous thrombosis, PE pulmonary embolus, SSSI superficial surgical site infection, DSSI deep surgical site infection, OSSI organ space surgical site infection
A total of 712 patients (56%) had resection for pancreatic adenocarcinoma. A subgroup analysis of postoperative outcomes among patients with pancreatic adenocarcinoma stratified by age is summarized in Table 4. Discharge to facility was the only postoperative outcome which differed between groups (p<0.001) with 52% of patients ≥ 81 discharged to facility after hospitalization.
Table 4.
Postoperative outcomes by age group in the subgroup of patients with pancreatic adenocarcinoma
| Age 65–74 (n = 432) |
Age 75–80 (n = 170) |
Age ≥ 81 (n = 110) |
p-value | |
|---|---|---|---|---|
| Composite morbidity | 187 (43.3) | 74 (45.4) | 53 (48.2) | 0.64 |
| Mortality | 11 (2.5) | 6 (3.5) | 4 (3.6) | 0.73 |
| Return to the operating room | 19 (4.4) | 5 (2.9) | 3 (2.7) | 0.57 |
| Discharge to facility | 49 (11.6) | 46 (27.4) | 56 (52.3) | < 0.001 |
| Unplanned readmission | 69 (16.0) | 24 (14.7) | 19 (17.3) | 0.84 |
| Pneumonia | 13 (3.0) | 9 (5.3) | 7 (6.4) | 0.19 |
| Reintubation | 18 (4.2) | 7 (4.1) | 9 (8.2) | 0.19 |
| Failure to wean ventilator | 10 (2.3) | 3 (1.8) | 4 (3.6) | 0.60 |
| MI | 5 (1.2) | 0 (0) | 2 (1.8) | 0.27 |
| Cardiac arrest | 9 (2.1) | 3 (1.8) | 1 (0.9) | 0.71 |
| Transfusion | 80 (18.5) | 27 (15.9) | 24 (21.8) | 0.45 |
| DVT | 11 (2.5) | 1 (0.6) | 2 (1.8) | 0.30 |
| PE | 4 (0.9) | 0 (0) | 0 (0) | 0.27 |
| Sepsis | 16 (3.7) | 11 (6.5) | 9 (8.2) | 0.10 |
| Septic shock | 9 (2.1) | 6 (3.5) | 6 (5.4) | 0.15 |
| SSSI | 36 (8.3) | 12 (7.1) | 6 (5.4) | 0.57 |
| DSSI | 3 (0.7) | 1 (0.6) | 1 (0.9) | 0.95 |
| OSSI | 43 (10.0) | 13 (7.6) | 13 (11.8) | 0.49 |
| Fascial dehiscence | 2 (0.5) | 1 (0.6) | 1 (0.9) | 0.85 |
| Renal insufficiency | 2 (0.5) | 0 (0) | 0 (0) | 0.52 |
| Acute renal failure | 3 (0.7) | 2 (1.2) | 1 (0.9) | 0.84 |
| Urinary tract infection | 12 (2.8) | 9 (5.3) | 4 (3.6) | 0.32 |
| Stroke | 2 (0.5) | 0 (0) | 0 (0) | 0.52 |
| Pancreatic fistula | 63 (14.6) | 25 (14.7) | 15 (13.6) | 0.96 |
| Length of stay (days), mean (SD) | 9.44 (8.3) | 9.86 (10.6) | 11.05 (7.8) | 0.23 |
MI myocardial infarction, DVT deep venous thrombosis, PE pulmonary embolus, SSSI superficial surgical site infection, DSSI deep surgical site infection, OSSI organ space surgical site infection
The presence of a single GSV was not associated with differences in composite morbidity, mortality, or unplanned readmissions (all p ≥ 0.08). Patients with at least one GSV were more likely to require a reoperation (3.1% vs 5.9%, p = 0.02) and to be discharged to a facility (11.6% vs 34.1%, p < 0.0001), Table 5.
Table 5.
Postoperative outcomes by presence of a geriatric-specific variable
| No GSV (n = 806) |
1+ GSV (n = 460) |
p-value | |
|---|---|---|---|
| Composite morbidity | 329 (40.8) | 207 (45.0) | 0.16 |
| Mortality | 14 (1.7) | 15 (3.3) | 0.08 |
| Return to the operating room | 25 (3.1) | 27 (5.9) | 0.02 |
| Discharge to facility | 93 (11.6) | 156 (34.1) | < 0.0001 |
| Unplanned readmission | 133 (16.5) | 76 (16.5) | 0.99 |
Adjusted associations between GSVs and mFI-5 with outcome measures
Multivariable models adjusting for preoperative ACS NSQIP estimated risk are summarized for mFI-5 (Table 6) and GSVs (Table 7). After adjusting for the ACS NSQIP estimated probability of morbidity, neither mFI-5 scores nor GSVs were associated with increased risk for postoperative morbidity (OR 1.06 [95% CI 0.49 – 2.31], p = 0.88 and OR 1.11 [95% CI 0.88 – 1.41], p = 0.38, respectively). After adjusting for the ACS NSQIP estimated probability of mortality, neither categorical GSVs nor mFI-5 scores were associated with increased risk for postoperative mortality (OR 1.60 [95% CI 0.76 – 3.39], p = 0.11 and OR 0.63 [95% CI 0.06 – 6.68], p = 0.99, respectively).
Table 6.
Unadjusted and adjusted performance of 5-factor modified frailty index scores for geriatric patients selected for pancreatic surgery
| Outcome | OR | 95% CI | p-value |
|---|---|---|---|
| Unadjusted | |||
| Composite morbidity | 2.22 | 1.07–4.64 | 0.03 |
| Mortality | 1.24 | 0.11–13.7 | 0.84 |
| Reoperation | 8.13 | 1.36–48.5 | 0.02 |
| Discharge to facility | 5.00 | 1.97–12.7 | 0.0002 |
| Unplanned readmission | 2.17 | 0.83–5.70 | 0.11 |
| Adjusted | |||
| Composite morbidity | 1.06 | 0.49–2.31 | 0.88 |
| Mortality | 0.63 | 0.06–6.68 | 0.99 |
| Reoperation | 3.62 | 0.57–23.1 | 0.18 |
| Discharge to facility | 1.83 | 0.69–4.87 | 0.23 |
| Unplanned readmission | 1.62 | 0.60–4.40 | 0.35 |
Table 7.
Unadjusted and adjusted performance of geriatric-specific variable model for geriatric patients selected for pancreatic surgery
| Outcome | OR | 95% CI | p-value |
|---|---|---|---|
| Unadjusted | |||
| Composite morbidity | 1.19 | 0.94–1.49 | 0.15 |
| Mortality | 1.91 | 0.91–3.99 | 0.09 |
| Reoperation | 1.95 | 1.12–3.40 | 0.02 |
| Discharge to facility | 4.05 | 3.00–5.46 | < 0.0001 |
| Unplanned readmission | 1.00 | 0.74–1.36 | 0.99 |
| Adjusted | |||
| Composite morbidity | 1.11 | 0.88–1.41 | 0.38 |
| Mortality | 1.60 | 0.76–3.39 | 0.11 |
| Reoperation | 1.81 | 1.03–3.16 | 0.04 |
| Discharge to facility | 3.95 | 2.91–5.38 | < 0.0001 |
| Unplanned readmission | 0.97 | 0.71–1.32 | 0.84 |
The presence of ≥ 1 GSV was associated with the need for reoperation and discharge to facility (OR 1.81 [95% CI 1.03 – 3.16], p = 0.04 and OR 3.95 [95% CI 2.91 – 5.38], p < 0.0001, respectively). There was no association between preoperative mFI-5 scores and reoperation or discharge to facility (both p ≥ 0.18). Neither the presence of a GSV nor the mFI-5 were associated with unplanned readmissions (both p ≥ 0.35). Associations between individual GSVs and primary and secondary outcome measures are summarized in Table 8. Each of the four individual GSVs were associated with discharge to facility (all p ≤ 0.004) and patients lived alone or at a facility were also more likely to have a reoperation or require an unplanned readmission (both p ≤ 0.001).
Table 8.
Adjusted performance of individual geriatric specific variables for geriatric patients selected for pancreatic surgery
| Outcome | OR | 95% CI | p-value |
|---|---|---|---|
| Lives home alone or at a facility | |||
| Composite morbidity | 1.02 | 0.77 – 1.34 | 0.91 |
| Mortality | 1.97 | 0.90 – 4.29 | 0.09 |
| Reoperation | 5.96 | 2.26 – 15.70 | 0.0003 |
| Discharge to facility | 2.53 | 1.88 – 3.41 | < 0.0001 |
| Unplanned readmission | 3.65 | 2.64 – 5.05 | < 0.0001 |
| Uses mobility aid | |||
| Composite morbidity | 1.19 | 0.87 – 1.65 | 0.28 |
| Mortality | 0.85 | 0.29 – 2.51 | 0.77 |
| Reoperation | 0.92 | 0.42 – 2.00 | 0.83 |
| Discharge to facility | 3.42 | 2.42 – 4.85 | < 0.0001 |
| Unplanned readmission | 1.15 | 0.76 – 1.72 | 0.51 |
| Fall history | |||
| Composite morbidity | 0.87 | 0.57 – 1.33 | 0.52 |
| Mortality | 0.84 | 0.20 – 3.57 | 0.81 |
| Reoperation | 0.61 | 0.19 – 2.02 | 0.42 |
| Discharge to facility | 2.16 | 1.37 – 3.41 | 0.001 |
| Unplanned readmission | 0.90 | 0.52 – 1.58 | 0.72 |
| Did not sign own consent | |||
| Composite morbidity | 0.90 | 0.31 – 2.59 | 0.84 |
| Mortality | 2.47 | 0.31 – 19.80 | 0.40 |
| Reoperation* | N/A | N/A | N/A |
| Discharge to facility | 5.00 | 1.66 – 15.05 | 0.004 |
| Unplanned readmission | 1.11 | 0.31 – 4.02 | 0.87 |
There were no patients who lacked capacity to sign their own consent and required a reoperation.
Discussion:
After adjustment for ACS NSQIP estimated risk of morbidity in geriatric patients selected for elective pancreatic resection, the presence of a single geriatric-specific variable is associated with reoperation and greater likelihood of discharge to a facility. In contrast, after adjusting for preoperative risk, the mFI-5 score is not independently associated with worse short-term postoperative outcomes. Crude and adjusted risk of discharge to a facility among geriatric patients with the presence of a GSV is particularly noteworthy; 34.1% of patients with a single GSV required discharge to a facility and discharge to a facility was particularly high among the elderly with 23% of patients between 75 and 80 discharged to a facility and 48% of patients 81 and greater discharged to a facility. After multivariable adjustment, the presence of a GSV was strongly associated with patient discharge to a facility (OR = 3.95). Importantly, while the proportion of patients with one or more GSV increased as age increased, the current models adjust for patient age as part of the ACS NSQIP estimated risk variable. Each individual GSV was associated with discharge to facility. Patients who resided home alone or at the facility, was the only GSV significantly associated with other outcome measures, including reoperation and unplanned readmission.
There were no associations between GSVs and 30-day morbidity and mortality in either unadjusted or adjusted analyses. The underlying implication underscores improvements in short-term morbidity and mortality among elderly patients selected for pancreatic resection. While recent registry data continues to demonstrate greater mortality rates among octogenarians and nonagenarians, pancreatic resections have been performed safely among the elderly in high volume centers.4,20,21 Interestingly, the presence of GSVs is associated with increased risk of reoperation. Factors associated with preoperative debility, such as use of mobility aid or recent history of falls, are linked with malnutrition and wound healing; these factors can significantly increase risk of a reoperation.22
Discharge to a facility has been previously described to be clinically important by having a significant impact on long-term outcomes. A prior study from our group evaluating the first two years of the Geriatric Surgery Research File data among all major abdominal surgery patients demonstrated similarly high discharge to facility rates.12 The current study focusing on pancreatic resection patients only, with two years of additional data, similarly identifies discharge to a facility as the major short-term negative outcome among geriatric patients selected for pancreatic surgery. Discharge to a facility among patients selected for pancreatic and major abdominal surgery has been associated with greater risk of 30- and 90-day readmission and 1-year mortality.23–25 While studies examining associations between GSVs and long-term outcomes are still lacking, a number of studies have evaluated readmission prevention strategies in post-pancreatectomy patients.26,27 The addition of these preoperative geriatric-specific variables to existing readmission models can potentially improve model performance which at present have modest performance (C-statistics between 0.66 and 0.76).
While GSVs were associated with two secondary outcome measures, mFI-5 scores were not associated with any primary or secondary outcome measures. A recently published study has similarly described poor prediction ability of mFI-5 among pancreatic surgery patients with average C-statistic of 0.51 in a national pancreatoduodenectomy cohort.28 One explanation is a lack of relevant variables in the mFI-5 composition. Only one of five variables in the score – dependent functional status – is related to a patient’s performance status. The remaining four variables – congestive heart failure, diabetes mellitus, hypertension, and chronic obstructive pulmonary disease – are measures of comorbid conditions and not frailty. Moreover, only 16 patients (1.3%) from the entire study cohort had dependent functional status, without statistically significant differences between patient age groups. With so few patients demonstrating findings of frailty using the mFI-5 score, the lack of significant associations between mFI-5 and outcome measures are not surprising.
A number of other frailty metrics have been proposed. One of the proposed metrics of frailty is sarcopenia. Sarcopenia has been estimated using variety of measures including grip strength and radiographic muscle measurements. Granular patient data are required to compute sarcopenia; these data are not available in registry studies. While not all sarcopenic patients are frail, and vice versa not all frail patients are sarcopenic, there is significant overlap between sarcopenia and frailty. A meta-analysis from 2019 evaluating sarcopenia in 7,176 patients (29 studies) with gastrointestinal cancers demonstrated association between sarcopenia and 30-40% increased risk of major and total complications after cancer resection, though much of the evidence was overall heterogenous and low quality.29 Another recently published meta-analysis of eighteen studies with 62,896 patients investigated sixteen different tools utilized to evaluate frailty and sarcopenia in vascular surgery patients. Similar concerns regarding the emphasis on comorbid conditions, rather than frailty measures, were described in this meta-analysis. Validated and reliable frailty measurement tools—including the clinical frailty scale, Edmonton frailty scale, and Fried criteria—were recommended for clinical practice and for future research over tools such as the mFI30 – all of these validated frailty measures require granular prospective data collection.
As in prior studies estimating the associations between independent variables and outcomes using the ACS NSQIP data, we used the ACS NSQIP estimated probabilities of morbidity and mortality to adjust for effects of possible confounders.12,31 Using these summary probabilities allows for statistical adjustment of the overall risk as estimated using variables included in the ACS NSQIP calculator, without constructing models with 20+ individual variables which can cause modeling problems with overfitting and multiple comparisons.32,33 Study limitations are similarly those inherent to ACS NSQIP designed studies including retrospective data collection, 30-day outcome period, and analyses that can test for association and not causation. Despite these limitations, data in this study quantifies the presence of GSVs in geriatric patients selected for pancreatic resection and describes significant association between the presence of GSVs and worse short-term outcome measures.
A recently published study summarizing over 38,000 patients included in the Geriatric Surgery Pilot Project, across ACS NSQIP procedure fields, demonstrated associations between geriatric-specific variables and four tested outcomes: pressure ulcers, delirium, functional decline, and new use of mobility aid.34 In addition, this study demonstrates associations between presence of any single GSV and greater risk of reoperation and discharge to a facility among elderly patients selected for pancreatic resection. GSVs are simple to screen during preoperative visits, are clinically meaningful, and have greater impact on short-term outcome measures than mFI-5 among pancreatic surgery patients. Incorporation of GSVs into preoperative risk calculation and the informed consent process will help improve estimation of post-operative risk and joint decision making.
Acknowledgment
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.
Funding:
This work was supported in part by National Cancer Institute awards: T32 CA163177 to Christopher L. Cramer and Courtney M. Lattimore, and 2 L30 CA220861-02A2 to Victor M. Zaydfudim.
Footnotes
This paper was presented as part of Society for University Surgeons program at the 17th Annual Academic Surgical Congress, February 1-3, 2022, Orlando, FL
References
- 1.Mohanty S, Rosenthal RA, Russell MM, Neuman MD, Ko CY, Esnaola NF. Optimal Perioperative Management of the Geriatric Patient: A Best Practices Guideline from the American College of Surgeons NSQIP and the American Geriatrics Society. J Am Coll Surg. 2016. May 1;222(5):930–47. [DOI] [PubMed] [Google Scholar]
- 2.Smith B, Smith G, Hurria A, Hortobagyi G, Buchholz T. Future of Cancer Incidence in the United States: Burdens Upon an Aging, Changing Nation. J Clin Oncol Off J Am Soc Clin Oncol. 2009. Apr 1;27:2758–65. [DOI] [PubMed] [Google Scholar]
- 3.Prashant S, Jonathan T, Mauricio S, James S, Peter D. Advanced age is a risk factor for post-operative complications and mortality after a pancreaticoduodenectomy: a meta-analysis and systematic review. HPB. 2012. Oct 1;14(10):649–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hue JJ, Bingmer K, Sugumar K, Ocuin LM, Rothermel LD, Winter JM, et al. Mortality and Survival Among Octogenarians with Localized Pancreatic Head Cancer: a National Cancer Database Analysis. J Gastrointest Surg Off J Soc Surg Aliment Tract. 2021. Oct;25(10):2582–92. [DOI] [PubMed] [Google Scholar]
- 5.Attard JA, Al-Sarireh B, Bhogal RH, Farrugia A, Fusai G, Harper S, et al. Short-term outcomes after pancreatoduodenectomy in octogenarians: multicentre case-control study. Br J Surg. 2021. Dec 17;109(1):89–95. [DOI] [PubMed] [Google Scholar]
- 6.Revenig LM, Canter DJ, Taylor MD, Tai C, Sweeney JF, Sarmiento JM, et al. Too Frail for Surgery? Initial Results of a Large Multidisciplinary Prospective Study Examining Preoperative Variables Predictive of Poor Surgical Outcomes. J Am Coll Surg. 2013. Oct 1;217(4):665–670.e1. [DOI] [PubMed] [Google Scholar]
- 7.Quero G, Pecorelli N, Paiella S, Fiorillo C, Petrone MC, Capretti G, et al. Pancreaticoduodenectomy in octogenarians: The importance of “biological age” on clinical outcomes. Surg Oncol. 2021. Nov 24;40:101688. [DOI] [PubMed] [Google Scholar]
- 8.Levi ST, Gough BL, Darcy CE, Petrelli NJ, Bennett JJ. Pancreatic resections: 30 and 90-day outcomes in octogenarians. Surg Oncol. 2021. Jun;37:101319. [DOI] [PubMed] [Google Scholar]
- 9.Eamer G, Al-Amoodi MJH, Holroyd-Leduc J, Rolfson DB, Warkentin LM, Khadaroo RG. Review of risk assessment tools to predict morbidity and mortality in elderly surgical patients. Am J Surg. 2018. Sep 1;216(3):585–94. [DOI] [PubMed] [Google Scholar]
- 10.Turrentine FE, Wang H, Simpson VB, Jones RS. Surgical Risk Factors, Morbidity, and Mortality in Elderly Patients. J Am Coll Surg. 2006. Dec 1;203(6):865–77. [DOI] [PubMed] [Google Scholar]
- 11.Robinson TN, Wu DS, Pointer LF, Dunn CL, Moss M. Preoperative Cognitive Dysfunction Is Related to Adverse Postoperative Outcomes in the Elderly. J Am Coll Surg. 2012. Jul 1;215(1):12–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Martin AN, Hoagland DL, Turrentine FE, Jones RS, Zaydfudim VM. Safety of Major Abdominal Operations in the Elderly: A Study of Geriatric-Specific Determinants of Health. World J Surg. 2020. Aug 1;44(8):2592–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chow WB, Rosenthal RA, Merkow RP, Ko CY, Esnaola NF. Optimal Preoperative Assessment of the Geriatric Surgical Patient: A Best Practices Guideline From the American College of Surgeons National Surgical Quality Improvement Program and the American Geriatrics Society. Surv Anesthesiol. 2013. Aug;57(4):179. [DOI] [PubMed] [Google Scholar]
- 14.Subramaniam S, Aalberg JJ, Soriano RP, Divino CM. New 5-Factor Modified Frailty Index Using American College of Surgeons NSQIP Data. J Am Coll Surg. 2018. Feb 1;226(2):173–181.e8. [DOI] [PubMed] [Google Scholar]
- 15.Subramaniam S, Aalberg JJ, Soriano RP, Divino CM. The 5-Factor Modified Frailty Index in the Geriatric Surgical Population. Am Surg. 2021. Sep 1;87(9):1420–5. [DOI] [PubMed] [Google Scholar]
- 16.Partridge JSL, Harari D, Dhesi JK. Frailty in the older surgical patient: a review. Age Ageing. 2012. Mar 1;41(2):142–7. [DOI] [PubMed] [Google Scholar]
- 17.Makary MA, Segev DL, Pronovost PJ, Syin D, Bandeen-Roche K, Patel P, et al. Frailty as a Predictor of Surgical Outcomes in Older Patients. J Am Coll Surg. 2010. Jun 1;210(6):901–8. [DOI] [PubMed] [Google Scholar]
- 18.Robinson TN, Eiseman B, Wallace JI, Church SD, McFann KK, Pfister SM, et al. Redefining Geriatric Preoperative Assessment Using Frailty, Disability and Co-Morbidity. Ann Surg. 2009. Sep;250(3):449–55. [DOI] [PubMed] [Google Scholar]
- 19.Robinson TN, Rosenthal RA. The ACS NSQIP Geriatric Surgery Pilot Project: Improving care for older surgical patients [Internet]. The Bulletin. 2014. [cited 2022 Jan 12]. Available from: https://bulletin.facs.org/2014/10/the-acs-nsqip-geriatric-surgery-pilot-project-improving-care-for-older-surgical-patients/ [PubMed] [Google Scholar]
- 20.Meltzer RS, Kooby DA, Switchenko JM, Datta J, Carpizo DR, Maithel SK, et al. Does Major Pancreatic Surgery Have Utility in Nonagenarians with Pancreas Cancer? Ann Surg Oncol. 2021. Apr;28(4):2265–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Makary MA, Winter JM, Cameron JL, Campbell KA, Chang D, Cunningham SC, et al. Pancreaticoduodenectomy in the very elderly. J Gastrointest Surg Off J Soc Surg Aliment Tract. 2006. Mar;10(3):347–56. [DOI] [PubMed] [Google Scholar]
- 22.Eachempati SR, Hydo LJ, Barie PS. Factors influencing the development of decubitus ulcers in critically ill surgical patients. Crit Care Med. 2001. Sep;29(9):1678–82. [DOI] [PubMed] [Google Scholar]
- 23.Paredes AZ, Hyer JM, Beal EW, Bagante F, Merath K, Mehta R, et al. Impact of skilled nursing facility quality on postoperative outcomes after pancreatic surgery. Surgery. 2019. Jul 1;166(1):1–7. [DOI] [PubMed] [Google Scholar]
- 24.Legner VJ, Massarweh NN, Symons RG, McCormick WC, Flum DR. The Significance of Discharge to Skilled Care After Abdominopelvic Surgery in Older Adults. Ann Surg. 2009. Feb;249(2):250–5. [DOI] [PubMed] [Google Scholar]
- 25.Hakkarainen TW, Arbabi S, Willis MM, Davidson GH, Flum DR. Outcomes of Patients Discharged to Skilled Nursing Facilities After Acute Care Hospitalizations. Ann Surg. 2016. Feb;263(2):280–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Valero V, Grimm JC, Kilic A, Lewis RL, Tosoian JJ, He J, et al. A novel risk scoring system reliably predicts readmission after pancreatectomy. J Am Coll Surg. 2015. Apr;220(4):701–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Turrentine FE, McMurry TL, Smolkin ME, Jones RS, Zaydfudim VM. Specialty-Specific Readmission Risk Models Outperform General Models in Estimating Hepatopancreatobiliary Surgery Readmission Risk. J Gastrointest Surg Off J Soc Surg Aliment Tract. 2021. Dec;25(12):3074–83. [DOI] [PubMed] [Google Scholar]
- 28.Al Abbas AI, Borrebach JD, Pitt HA, Bellon J, Hogg ME, Zeh HJ, et al. Development of a Novel Pancreatoduodenectomy-Specific Risk Calculator: an Analysis of 10,000 Patients. J Gastrointest Surg Off J Soc Surg Aliment Tract. 2021. Jun;25(6):1503–11. [DOI] [PubMed] [Google Scholar]
- 29.Simonsen C, de Heer P, Bjerre ED, Suetta C, Hojman P, Pedersen BK, et al. Sarcopenia and Postoperative Complication Risk in Gastrointestinal Surgical Oncology: A Meta-analysis. Ann Surg. 2018. Jul;268(1):58–69. [DOI] [PubMed] [Google Scholar]
- 30.Houghton JSM, Nickinson ATO, Morton AJ, Nduwayo S, Pepper CJ, Rayt HS, et al. Frailty Factors and Outcomes in Vascular Surgery Patients: A Systematic Review and Meta-analysis. Ann Surg. 2020. Aug;272(2):266–76. [DOI] [PubMed] [Google Scholar]
- 31.Elwood NR, Martin AN, Turrentine FE, Jones RS, Zaydfudim VM. The negative effect of perioperative red blood cell transfusion on morbidity and mortality after major abdominal operations. Am J Surg. 2018. Sep;216(3):487–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cohen ME, Ko CY, Bilimoria KY, Zhou L, Huffman K, Wang X, et al. Optimizing ACS NSQIP modeling for evaluation of surgical quality and risk: patient risk adjustment, procedure mix adjustment, shrinkage adjustment, and surgical focus. J Am Coll Surg. 2013. Aug;217(2):336–346.e1. [DOI] [PubMed] [Google Scholar]
- 33.Liu Y, Cohen ME, Hall BL, Ko CY, Bilimoria KY. Evaluation and Enhancement of Calibration in the American College of Surgeons NSQIP Surgical Risk Calculator. J Am Coll Surg. 2016. Aug;223(2):231–9. [DOI] [PubMed] [Google Scholar]
- 34.Hornor MA, Ma M, Zhou L, Cohen ME, Rosenthal RA, Russell MM, et al. Enhancing the American College of Surgeons NSQIP Surgical Risk Calculator to Predict Geriatric Outcomes. J Am Coll Surg. 2020. Jan;230(1):88–100.e1. [DOI] [PubMed] [Google Scholar]