Blood Loss from Robotic Assisted Hysterectomy

Introduction

Approximately 600,000 hysterectomies are performed annually in the United States, making the hysterectomy the most commonly performed gynecologic surgery. ¹ While abdominal hysterectomy via laparotomy continues to be the most common approach, the advent of laparoscopic hysterectomy in 1989 began a shift in practice toward minimally invasive techniques. ² By 2011, a reported one-third of hysterectomies performed for endometrial cancer utilized a minimally invasive technique. ³ In 2005, the U.S. Food and Drug Administration approved the daVinci Surgical System (Intuitive Surgical Inc., Sunnyvale, CA) for gynecologic surgery. This quickened the shift toward laparoscopy, due in part to the assertion that the robotic platform is easier to master than conventional “straight-stick” laparoscopy. ⁴ However, little data exist suggesting perioperative benefit of the daVinci system over laparoscopy for hysterectomy performed for benign conditions.^5,6

One major complication of hysterectomy is acute blood loss anemia. However, it has been reported that blood loss during laparoscopic hysterectomy is significantly less than when the hysterectomy is performed via laparotomy. ⁷ This has been suggested to be true of the daVinci robotic assisted procedures as well, thought to be enhanced due to the use of high definition, three-dimensional optics and magnification that allow better identification of blood vessels, allowing the surgeon to seal these vessels before cutting them.

Following robotic assisted hysterectomy, the current authors have observed in their patients a drop in postoperative hemoglobin (Hb) despite minimal intraoperative estimated blood loss (EBL). The current authors sought to compare postoperative decrease in Hb level with EBL at the time of surgery, as well as in the magnitude of the decrease in Hb level in patients undergoing total abdominal hysterectomy (TAH) or robotic assisted hysterectomy (RAH). An attempt was also made to identify factors contributing to the postoperative decrease in Hb.

Materials and Methods

After receiving institutional review board (IRB) approval (IRB Protocol #23130), records of 270 patients who underwent hysterectomy at an urban tertiary care center between 2011 (the initiation of the robotic surgery program at the institution) and 2015 were reviewed retrospectively. Included were all robotic assisted and total abdominal hysterectomies performed during this time period by 2 experienced surgeons in the institution's division of gynecologic oncology, with the exception of hysterectomies performed during ovarian cancer debulking. Cases were identified using an institutional surgical case log and data were abstracted from the electronic medical records.

Data were abstracted and entered into an electronic spreadsheet by 1 of the investigators (D.S.) and were audited by 2 other investigators (E.H. and J.S.F.). Data were shown as the median (interquartile range [IQR], defined as the 25th–75th percentiles) for continuous variables and the count and percentage for categorical variables. Both absolute change and relative percent change in Hb level were examined as primary endpoints. The preoperative Hb was obtained within 14 days of the surgery. The postoperative Hb was obtained ∼12 hours after completion of the surgery. The data were summarized overall for all the patients as well as by surgery type.

The χ² or Fisher's exact test was used when comparing categorical variables, and the Wilcoxon rank-sum test was used to compare continuous variables between the 2 surgery groups. Multiple linear regression analyses were performed to evaluate the associations between various factors, such as patient demographics and baseline disease characteristics (e.g., age, race, body mass index [BMI], preoperative Hb level, fluid balance, and uterine weight) and the primary outcome variables. All 2-way interaction terms between the type of surgery (RAH versus TAH) and the other potential predictors were considered for possible inclusion in the final regression model. The variables were selected for the final regression model using a stepwise method with the significance level for retention of a variable in a model being <0.05. Data transformation (e.g., log or square-root) was considered when a variable on the original scale was not approximately normally distributed. p-Values <0.05 were considered statistically significant. The data were analyzed (D.Y.) using SAS version 9.3 (SAS Institute Inc., Cary, NC).

Results

A total of 270 hysterectomies were studied, with 90 robotic assisted total laparoscopic hysterectomies (RAHs) and 180 TAHs. Empirical comparisons of the two groups showed no clinically significant differences in age, race, parity, gravidity, or BMI (Table 1). The uterine weight was greater in patients who underwent TAH, compared to those who underwent RAH (median [IQR]: 169 g [95–398 g] versus 120 g (86–164 g]). Operative time was longer for RAH (median [IQR]: 201 minutes [163.5–257.5 minutes] versus 148 minutes [127.0–190.0 minutes] for TAH).

While histories of prior abdominal surgery was similar between the two groups, a larger proportion (45.5%) of women in the TAH group had undergone prior laparoscopic surgery, compared to those in the RAH group (31.3%). Patients underwent concurrent pelvic lymph node dissection more frequently in the RAH group (38% versus 21%), but the percentage of patients undergoing para-aortic lymph-node dissection was the same in both groups (6.7%). A similar proportion of patients received preoperative heparin in both groups (59%). Approximately half of all patients underwent other concurrent procedures at the time of hysterectomy, and the proportion of patients who underwent concurrent procedures at the time of TAH (51%) was similar to that of patients who underwent RAH (43%). Uterine and cervical cancer was the indication for surgery in 75% of the patients who underwent RAH and in 47.2% of those who underwent TAH.

While the EBL was higher in patients who underwent TAH, compared to RAH (median [IQR]: 300 mL [150–500 mL] versus 100 mL [50–150 mL]; p < 0.0001), the postoperative day 1 Hb was similar (median [IQR]: 10.8 g/dL [9.7–11.5 g/dL] versus 10.4 g/dL [9.1–11.3 g/dL]). The median (IQR) preoperative Hb was 13.1 g/dL (12.3–14.0 g/dL) for patients who underwent RAH and 12.2 g/dL (10.8–10.3 g/dL) for those who underwent TAH. The median (IQR) decrease in Hb was higher in patients who underwent RAH, compared to those who underwent TAH (2.3 g/dL [1.5–3.2 g/dL] versus 1.9 g/dL [1.2–2.4 g/dL]; p < 0.001). Seventeen (9.4%) patients in the TAH group received blood transfusions, but only 1 (1.1%) did in the RAH group. The median number of units transfused was 2 in both groups. When patients who received blood transfusions were excluded from the analysis, the patients in the RAH group still had a higher decrease in Hb than those in the TAH group (median [IQR]: 2.3 g/dL [1.5–3.1 g/dL] for RAH versus 1.9 g/dL [1.2–2.4 g/dL] for laparotomy; p = 0.001).

The estimated fluid balance between the time of surgery and the morning of the first postoperative day was analyzed. The median (IQR) fluid balance in the RAH group was positive 445 mL (–412 to 1069 mL). The fluid balance in the TAH group was negative 455 mL (–1612 to 502) mL. Multiple regression analysis showed an association of fluid balance with change in Hb. However, after controlling for BMI and preoperative Hb level, the association was no longer statistically significant.

The significant predictors of Hb change using multiple regression modeling were the type of surgery performed, preoperative Hb, and BMI. RAH, higher preoperative Hb, and lower BMI were associated with a larger drop in Hb level after hysterectomy. In patients who underwent RAH the percent of decrease in Hb level was higher among those who underwent concurrent pelvic lymph-node dissection (median [IQR]: 20.4% [14.2%–25.9%] versus 17.1% [11.1%–20.7%]; p = 0.05). Linear regression modeling, using type of surgery (RAH versus TAH), preoperative Hb, fluid balance, and whether or not pelvic lymph-node dissection was performed as potential predictors and/or adjustment variables indicated that pelvic lymph-node dissection was a strong predictor of the change in Hb (0.6 larger decrease if pelvic lymph-node dissection was performed; p < 0.001), along with preoperative Hb (slope in preoperative Hb: 0.27; p < 0.0001) and fluid balance (slope in fluid balance: 0.0001; p = 0.012).

Discussion

In this retrospective review of hysterectomies performed at a gynecologic oncology service, the current authors sought to investigate the clinical observation of a disconnect between EBL and postoperative Hb level, particularly among robotic-assisted cases. It was found that the data supported the observation that RAH was associated with a decrease in postoperative Hb greater than would be expected given the EBL. Additionally, it was found that lymphadenectomy performed at the time of RAH was an independent predictor of a larger change in postoperative Hb level. However, only 1 patient in the RAH group received a blood transfusion.

It has been stated that the estimation of surgical blood loss frequently underestimates true blood loss. ⁷ It is possible that underestimation of blood loss accounted for some of the unexplained decreases in Hb observed in the RAH patients. Fluid balance is another factor often considered when evaluating postoperative anemia. The RAH patients had a positive fluid balance on postoperative day 1. It could be assumed that hemodilution was responsible for the observed change in Hb levels. However, experimental evidence shows that, while an intravenous fluid bolus followed by maintenance fluids results in a decrease in Hb, the change is transient and the Hb level returns to baseline over a period of a few hours. ⁸ The current statistical analysis showed that fluid balance was not associated with the change in Hb when controlling for BMI and preoperative Hb level.

The concurrent performance of a pelvic lymph-node dissection was associated with a higher decrease in postoperative Hb. For the RAH cases, lymphadenectomy was independently predictive of this decrease. During laparoscopic surgery, intra-abdominal pressure is maintained at ∼15 mm Hg throughout the procedure. This amount of pressure can tamponade small veins in the pelvis. The current authors postulate that the most likely mechanism for the observed decrease in Hb in RAH patients, compared to the EBL was occult bleeding that occurred after the surgery was completed and the pneumoperitoneum was released.

Weaknesses of the current study are that the data were only procured from 2 experienced gynecologic oncologists at one urban tertiary-care institution. Most of the operations were not performed for benign indications, and one-half included concurrent procedures, such as lymph-node dissection, which might have increased the magnitude of the Hb drop. These factors could limit generalizability of the study outcomes to benign gynecology, as increased blood loss may be less problematic in patients without malignancy. ⁹

The current study was also retrospective, and, as such, patients were not randomized to one surgical modality or the other. Differences in BMI and concurrent procedures between the two groups point to inherent differences that could have increased the differences between the groups. Future researchers may attempt to match these parameters in order to minimize the amount of “apples-to-oranges” comparison being made.

A strength of the study is that the same surgeons performed both the TAH and RAH procedures, thus allowing for a relative comparison between the two surgical approaches. The results were similar even after excluding patients who received blood transfusion from the analysis. This demonstrates that the results were not skewed by a few cases of extreme blood loss.

Conclusions

Minimally invasive surgery has several documented benefits over laparotomy and is the preferred surgical approach for a significant proportion of gynecologic oncology patients. The robotic surgical platform has been adopted by a majority of gynecologic oncologists in the United States. ¹⁰ In this study, the low EBL did not correlate with the observed change in postoperative Hb, more so in robotic cases, compared to laparotomy with a higher decrease in Hb in RAH than TAH cases. This was despite the fact that the patients who underwent TAH had larger uteri and larger BMIs than patients who underwent RAH. Lymphadenectomy was significantly associated with the drop in Hb; this knowledge could also be used to modify practice.