Back to 2018 Program and Abstracts
TIMING AND OUTCOMES OF ABDOMINAL SURGERY IN NEUTROPENIC PATIENTS
Joshua S. Jolissaint*1, Maya Harary2, Lily V. Saadat1, Arin L. Madenci1, Ali Tavakkoli1
1Surgery, Brigham and Women's Hospital, Boston, MA; 2Harvard Medical School, Boston, MA
Introduction: Performing surgery in neutropenic patients is feared and with increasing utilization of chemotherapy agents, more frequently encountered. Understanding of the outcomes of abdominal surgery in this patient cohort is however limited to small retrospective series and case reports. We sought to analyze and report on the morbidity and mortality of abdominal surgery in neutropenic patients.
Methods: We conducted a single-institution retrospective review of patients (2001-2017) who were neutropenic (absolute neutrophil count [ANC] <1500 cells/mL) in the 24-hours prior to their index abdominal operation. Patients were characterized based on National Cancer Institute definitions as having mild (ANC 1000≤1500), moderate (ANC 500≤1000) or severe (ANC ≤500) neutropenia. Primary and secondary outcomes were 30-day mortality and morbidity, respectively. Morbidity was defined as ≥ 1 complication based on a subset of the National Surgical Quality Improvement Program (NSQIP) tracked outcomes. Chi-squared test and two-tailed Student's t-test were employed (non-parametric tests used when appropriate). To determine the optimal threshold of ANC to discriminate mortality, we maximized the Youden index (J), a statistic that equally weights sensitivity and specificity.
Results: Among 209 neutropenic patients, mortality was 12% (26/209) and morbidity 64% (133/209). ANC ≤500 (14/26 (54%) vs. 34/183 (19%), P<0.01), urgent/emergent operations (26/26 (100%) vs. 125/183 (68%), P<0.01), and perforated viscus (10/26 (38%) vs. 29/183 (16%), P<0.01) were each significantly associated with mortality (Table 1). Perforated viscus (37/133 (28%) vs. 2/76 (3%), P<0.01), open operations (109/133 (82%) vs. 44/76 (58%), P<0.01), transfer from an outside hospital (32/133 (24%) vs. 7/76 (9%) P<0.01), and days from admission to operation (median=1 day, IQR=0-6 days vs. median=1 day, IQR=0-3 days, P<0.01) were significantly associated with morbidity (Table 2). Based on receiver operating characteristic (ROC) curves for mortality, an ANC threshold of 350 provided the best discrimination for mortality.
Conclusion: Factors including low ANC, urgent/emergent operations, and an operative diagnosis of perforated viscus significantly increased patients' likelihood of both mortality and morbidity. An ANC of 350 may be a better threshold for defining "severe" disease in the surgical population. Both time from admission to operation and transfer from an outside hospital were associated with morbidity. Patients may experience operative delays in transferring to a quaternary care facility or in an attempt to optimize their fitness for surgery, however with the potential for adverse postoperative outcomes. As such, these data may support earlier intervention for patients with obvious surgical pathology.
Table 1. Factors associated with operative mortality among 209 neutropenic patients
| Variable | Mortality | No Mortality | P |
| Overall | 26 (12) | 183 (88) | |
| Age | 58.5 (49-72) | 56 (45-67) | 0.42 |
| Female | 13 (50) | 96 (52) | 0.81 |
| Non-elective | 26 (100) | 125 (68) | <0.01 |
| Open | 23 (88) | 130 (71) | 0.06 |
| ANC | 0.3 (0.2-1.0) | 1.1 (0.6-1.4) | <0.01 |
| ANC 0-500 | 14 (54) | 34 (19) | <0.01a |
| ANC 500-1000 | 5 (19) | 49 (27) | |
| ANC >1000 | 7 (27) | 100 (55) | |
| Etiology | | | 0.02 |
| Chemotherapy | 14 (54) | 88 (48) | |
| Idiopathic | 1 (4) | 44 (24) | |
| Immunosuppression | 6 (23) | 35 (19) | |
| Progression of Primary Disease | 2 (8) | 12 (7) | |
| Sepsis | 3 (12) | 4 (2) | |
| Small Bowel Obstruction | 4 (15) | 34 (19) | >0.99 |
| Perforation | 10 (38) | 29 (16) | 0.01 |
| Outside Transfer | 8 (31) | 31 (17) | 0.09 |
| GCSF | 8 (31) | 29 (16) | 0.09 |
| Days to Operation | 1 (0-7) | 1 (0-5) | 0.73 |
Number (%) or median (interquartile range).
aCochran-Armitage trend test. ANC=absolute neutrophil count, GCSF=granulocyte colony stimulating factor
Table 2. Factors associated with operative morbidity among 209 neutropenic patients
| Variable | Morbidity | No Morbidity | P |
| Overall | 133 (64) | 74 (36) | |
| Age | 58 (47-68) | 54.5 (43-66) | 0.15 |
| Female | 65 (49) | 44 (58) | 0.25 |
| Non-Elective | 111 (83) | 40 (53) | <0.01 |
| Open | 109 (82) | 44 (58) | <0.01 |
| ANC | 0.9 (0.4-1.3) | 1.1 (0.7-1.4) | 0.02 |
| ANC 0-500 | 35 (27) | 13 (17) | 0.11a |
| ANC 500-1000 | 33 (25) | 21 (27) | |
| ANC >1000 | 63 (48) | 44 (56) | |
| Etiology | | | 0.04 |
| Chemotherapy | 63 (48) | 39 (50) | |
| Idiopathic | 22 (17) | 23 (29) | |
| Immunosuppression | 30 (23) | 11 (14) | |
| Progression of Primary Disease | 9 (7) | 5 (6) | |
| Sepsis | 7 (5) | 0 (0) | |
| Small Bowel Obstruction | 21 (16) | 17 (22) | 0.27 |
| Perforation | 37 (28) | 2 (3) | <0.01 |
| Outside Transfer | 32 (24) | 7 (9) | <0.01 |
| GCSF | 30 (23) | 7 (9) | 0.01 |
| Days to Operation | 1 (0-6) | 1 (0-3) | <0.01 |
Number (%) or median (interquartile range).
aCochran-Armitage trend test. ANC=absolute neutrophil count, GCSF=granulocyte colony stimulating factor
Back to 2018 Program and Abstracts
