Extended Length of Stay After Robotic-Assisted Hysterectomy: Association with Uterine Weight and Other Risk Factors

Abstract

Background: Robotic-assisted hysterectomy (RH) has the same clinical benefits as laparoscopy and offers surgeons additional benefits. However, RH-related costs are high and its clinical benefits have not been shown to be better than those achieved with laparoscopy. A key cost factor is the length of (hospital) stay (LOS). Objective: The aim of this study was to identify the relationship between uterine weight and LOS following RH and potential risk factors for extended LOS. Materials and Methods: This study involved a retrospective cohort of all RHs performed in a midwestern tertiary-care teaching hospital and its suburban affiliates, from January 2011 to December 2012. Data were collected using Current Procedural Terminology codes. The current authors examined if any of several variables were associated with uterine weight, using Spearman's correlation for continuous variables and Wilcoxon's rank sum or Kruskal-Wallis test for categorical variables. Comparison of variables between patients who did and who did not have a LOS>1 day was performed using a Wilcoxon rank sum test for continuous variables and a Chi-square or Fisher's exact test for categorical variables. Those that were associated with both uterine weight and LOS were considered as potential confounders and were included in the logistic regression model. The adjusted odds ratio (OR) for a 100-g increase in uterine weight was calculated. Results: Of 239 patients who underwent RH, 48 (20%) had a LOS>1 day. Uterine weight was significantly greater among patients with LOS>1 day (483 g versus 337 g; p=0.008). Patients who had LOS>1 day had a greater estimated blood loss (EBL; means: 178 mL versus 95 mL; p=0.006) and a significantly longer procedure duration (means: 236 minutes versus 168 minutes; p<0.005), compared to patients with LOS=1. In addition, patients with LOS>1 day had higher baseline pain scores (4.5 versus 3.2; p=0.003). Number of ports and oophorectomy were both significantly associated with LOS. For a 100-g increase in uterine weight, there was 1.12 times the odds of having LOS>1 day (OR=1.12; 95% confidence interval: 1.02, 1.21). After controlling for procedure duration, EBL, number of ports, transfusion rates and oophorectomy, the adjusted OR (aOR) was not significant (aOR=1.0; CI: 0.89, 1.12). Conclusions: When adjusted for procedure duration, EBL, number of ports, transfusion rates, and oophorectomy, uterine weight was not associated with LOS after RH. Larger studies are needed to explore further the associations among procedure duration, EBL, baseline pain scores, oophorectomy, and number of ports used with prolonged LOS after RH. Pain management seemed to be the single greatest indication for increased LOS after RH. (J GYNECOL SURG 32:19)

Introduction

Robotic-assisted hysterectomy (RH) was approved by the U.S. Food and Drug Administration as a minimally invasive approach to hysterectomy in 2005. The adoption of RH in clinical practice was rapid, and, in only 3 years, the procedure accounted for almost 10% of hysterectomies performed in the United States.¹ Robotic surgery offers the same clinical benefits of laparoscopy, including smaller surgical incisions, less postoperative pain, reduced intraoperative blood loss, and shorter recovery and hospitalization time. In addition, RH provides improved dexterity, more-precise movements, tremor reduction, better visualization of the operative field, a shorter learning curve, and less surgeon fatigue at the cost of tactile feedback.² For well-trained surgeons, these advantages could enable the performance of complicated surgeries—that might otherwise require laparotomies—to be completed via a minimally invasive procedure.^1,3

The major disadvantage of robotic surgery is the high costs of the robotic system, including costs of purchase, maintenance, and instruments. Many recent large comparative studies have concluded that robotic surgery is not associated with improved effectiveness or safety, compared to laparoscopy; morbidity profiles and clinical outcomes were similar with the only difference being a substantial increase in cost of robotic surgery.^1,3–5 In addition to the cost of equipment and the cost of operative time, the cost of the length of (hospital) stay (LOS) represents an important part of the overall cost of robotic surgery. The LOS is an important modifiable factor that can offer the possibility of decreasing the cost of robotic surgery by decreasing the number of inpatient hospital days.² Identifying factors associated with an increased LOS after RH is important for guiding future efforts toward improving quality of care and decreasing costs.

The current authors wanted to study risk factors associated with prolonged LOS after RH. Specific interest was focused on uterine weight as a potential factor associated with LOS in patients undergoing RH, as understanding the relationship between uterine weight and LOS could improve anticipation of a prolonged LOS.

Materials and Methods

This was a retrospective cohort of all patients who had RH at the Henry Ford Hospital, in Detroit, MI and its suburban affiliates, between January 2011 and December 2012. Institutional review board approval was received prior to data collection. All laparoscopic hysterectomies performed between 2011 and 2012 were identified using Current Procedural Terminology (CPT) codes. A manual chart review was performed for all cases, and cases for which conventional laparoscopic approach was used were excluded. Gynecologic oncology cases and primary urogynecologic cases were excluded. There is no CPT code for conversion from laparoscopic hysterectomy to an open procedure. Converted cases were likely to be recorded with an open CPT code; thus it was not possible to identify converted cases. Two hundred and forty one patients were finally included in the study. A systematic chart review was then conducted for each medical record. Two patients were missing their admission and discharge information; therefore 239 women of the original 241 were included in the analyses.

The outcome of interest was if each patient's LOS exceeded 1 day (LOS>1 day). The focus was on the role that uterine weight—defined in g based on the pathology report—played in LOS. Additional variables included in the analyses were: age at surgery; body mass index (BMI; in kg/m²); procedure duration (cut-to-close time according to the surgical report); estimated blood loss (EBL; in mL as recorded in the anesthesiology report); number of laparoscopic ports used (3, 4, or 5+; per the surgeon's note); postoperative complications (any or none, as documented in the medical records for the periods of hospital stay); blood transfusions (any or none during surgery or admission); readmission within 90 days (“yes” or “no”) and/or reoperation within 90 days (“yes” or “no”); history of depression or anxiety (“yes” or no, as documented in the medical records); baseline pain scores (as documented by the nursing staff in preoperative area); and addition of oophorectomy (none, unilateral, or bilateral per the surgeons' notes) or other procedures (any or none) during the RH.

Percentages and means with standard deviations were used to describe the study population. The current authors examined whether any of the independent variables were associated with uterine weight, using a Spearman correlation for continuous variables and a Wilcoxon rank sum or Kruskal-Wallis test for the categorical variables. Wilcoxon rank sum for continuous variables and Chi-square and Fisher's exact tests for categorical variables were used to compare variables between patients who did and did not have LOS>1 day.

The odds ratio (OR) for the association between uterine weight and having a LOS>1 day was calculated using a logistic regression model. The variables that were associated with both uterine weight and LOS were considered as potential confounders and were included in the adjusted logistic regression model. The adjusted OR (aOR) and 95% confidence interval (CI) for a 100-g increase in uterine weight were calculated from the estimates of the final logistic regression model. All analyses were performed using SAS version 9.4 (Cary, NC).

Results

Twenty percent of the patients (n=48) had a LOS>1 day. The mean uterine weight was 365.8 g (standard deviation [SD]=340.6 g). The mean age of the population was 45.0 years and the mean BMI 35.9 kg/m² (Table 1). The average procedure duration was ∼ 182 minutes with a mean EBL of 112.2 mL. Approximately 7.1% of patients had postoperative complications, and 2.9% needed blood transfusions. The majority of patients did not have oophorectomy at the time of hysterectomy (68.4%). Additional procedures performed at time of RH included salpingo-oophorectomy, resection of endometriosis, vaginal urogynecologic procedures (incontinence surgery, anterior/posterior colporrhaphy), lysis of adhesions, ureterolysis, appendectomy, and umbilical hernia repair. Gynecologic indications for the hysterectomy included leiomyoma, endometriosis, abnormal uterine bleeding, pelvic pain, abnormal endometrial or cervical pathology, and pelvic-organ prolapse.

Table 1.

Patient Characteristics & Surgical Factors (N=239)

Characteristics	Mean (SD) or n (%)
Age (in years)	45.0 (6.2)
BMI	35.9 (12.6)
Parity	1.8 (1.5)
Uterine weight	365.8 (340.6)
Procedure duration	182.0 (79.8)
EBL	112.2 (121.8)
Postoperative complications	17 (7.1%)
Transfusion rate	7 (2.9%)
Readmission	13 (5.5%)
Reoperation	7 (2.9%)
Any depression/anxiety	46 (19.3%)
Pain score	3.5 (3.1)
Oophorectomy
None	160 (68.4%)
Unilateral	33 (14.1%)
Bilateral	41 (17.5%)
Any other procedure	130 (54.3%)

SD, standard deviation; BMI, body mass index; EBL, estimated blood loss.

Age, BMI, procedure duration, and EBL were positively correlated with uterine weight (Table 2A). Cases with higher number of ports used and patients who received blood transfusions were both associated with larger uterine weight (Table 2B). However, patients who had postoperative complications had smaller uterine weights, compared to patients who did not have any complications.

Table 2A.

Spearman Correlations Between Uterine Weight & Continuous Variables

Characteristics	Spearman correlation with uterine weight	p-Value
Age (in years)	0.14	0.03
BMI	0.33	<0.001
Procedure duration (in minutes)	0.46	<0.001
EBL (in mL)	0.37	<0.001
Pain score	−0.06	0.35

Table 2B.

Mean (SD) Uterine Weight by Subgroups

Characteristics	n	Mean uterine weight (SD)	p-Value^*
Number of ports
3 ports	21	203.5 (145.3)	0.006
4 ports	157	335.1 (262.9)
5 or 6 ports	59	509.7 (496.7)
Transfusion
Yes	7	544.7 (362.9)	0.10
No	231	360.4 (339.3)
Any post-operative complication
Yes	16	207.4 (159.0)	0.03
No	222	377.3 (347.5)
Concomitant procedure^*
Yes	130	355.7 (366.8)	0.15
No	108	378.1 (307.4)
Oophorectomy
None	159	401.7 (361.8)	0.03
Unilateral	33	306.5 (229.5)
Bilateral	41	305.4 (329.1)
Depression/anxiety
Yes	46	311.8 (241.5)	0.36
No	191	379.4 (360.5)

p-Value for Wilcoxon rank sum (2 groups) or Kruskal–Wallis (>2 groups).

BMI, body mass index; EBL, estimated blood loss; SD, standard deviation.

Uterine weight was significantly associated with LOS in the bivariate analysis (Table 3A). Patients with LOS>1 day had larger uterine weights, compared to women with LOS≤1 (means: 483 g versus 336 g; p=0.007). Age and BMI did not differ between the LOS groups. Patients with LOS>1 day had a higher mean EBL (178 mL versus 95 mL; p=0.006) and longer procedure duration (236 minutes versus 168 minutes; p<0.005), compared to women with≤1 day LOS (Table 3A). In addition, patients with extended LOS had a higher baseline pain score (4.5 versus 3.2; p=0.003).

Table 3A.

Means (SD) of Characteristics by LOS (N=239)

Characteristic	LOS≤1	LOS>1	p-Value^*
Uterine weight (in g)	336.1 (296.3)	483.4 (462.9)	0.007
Age (in years)	45.1 (6.3)	44.6 (5.7)	0.67
BMI	35.3 (12.3)	38.4 (13.6)	0.14
Procedure duration in minutes	168.4 (59.9)	235.8 (118.4)	<0.005
EBL (in mL)	95.2 (80.7)	178.4 (206.2)	0.006
Pain score	3.2 (3.1)	4.5 (2.8)	0.003

The percentage of patients who had a LOS>1 day varied with the number of ports (Table 3B): 3 ports (23.8%), 4 ports (13.9%), or 5+ ports (35.6%). Patients who had blood transfusions were more likely to have stayed>1 day compared to patients who did not have blood transfusions (85.7% versus 18.1%, p<0.05). Having an oophorectomy was also associated with having a LOS>1 day: 16.3% of women who did not have oophorectomies; 27.3% of patients who had unilateral oophorectomies; and 29.3% of those who had bilateral oophorectomies had a LOS>1 day. Having postoperative complications, concomitant procedures, readmissions, or history of depression were not associated with LOS.

Table 3B.

Percentages of Subgroups by LOS

Characteristics	n	% LOS≤1	% LOS>1	p-Value†
Number of ports				0.002
3 ports	21	16 (76.2%)	5 (23.8%)
4 ports	158	136 (86.1%)	22 (13.9%)
5 or 6 ports	59	38 (64.4%)	21 (35.6%)
Transfusion				<0.001
Yes	7	1 (14.3%)	6 (85.7%)
No	232	190 (81.9%)	42 (18.1%)
Oophorectomy				0.04
None	160	134 (83.8%)	26 (16.3%)
Unilateral	33	24 (72.7%)	9 (27.3%)
Bilateral	41	29 (70.7%)	12 (29.3%)
Any postoperative complications				0.12
Yes	17	11 (64.7%)	6 (35.3%)
No	222	180 (81.1%)	42 (18.9%)
Concomitant procedure				0.54
Yes	130	102 (78.5%)	28 (21.5%)
No	109	89 (81.7%)	20 (18.4%)
Readmission				0.30
Yes	13	9 (69.2%)	4 (30.8%)
No	225	181 (80.4%)	44 (19.6%)
Depression/anxiety				0.23
Yes	46	34 (73.9%)	12 (26.1%)
No	192	157 (81.8%)	35 (18.2%)

p-Values are for Wilcoxon rank-sum test.

†p-Values are for Fisher's exact test or Chi-square test.

SD, standard deviation, LOS, length of (hospital) stay; BMI, body mass index; EBL, estimated blood loss.

Uterine weight was weakly associated with the odds of having a LOS>1 day (OR=1.12 for a 100-g increase in uterine size; 95% CI: 1.02, 1.21; Table 4). After controlling for procedure duration, EBL, number of ports, transfusion rates, and oophorectomy, the aOR was not significant (aOR=1.0; 95% CI: 0.89, 1.12) for a 100-g increase in uterine size).

Table 4.

Crude & Adjusted Odds Ratios Between Uterine Weight & LOS

Crude or adjusted	OR^* (95% CI)
Unadjusted	1.12 (1.02, 1.21)
Adjusted^a	1.0 (0.89, 1.12)

Based on a logistic regression model, adjusted for procedure duration, estimated blood loss, number of ports, transfusion, and oophorectomy.

OR is for 100-g increase in uterine weight.

LOS, length of (hospital) stay; OR, odds ratio; CI, confidence interval.

The indications for extended hospital stay in patients with LOS>1 day were documented as follows: pain control (n=19); postoperative ileus and nausea/vomiting (n=6); urinary retention (n=2); prolonged surgery (n=3; 2 requiring postoperative admission to the intensive care unit [ICU]); postoperative fever (n=2; atelectasis, pneumonia); chest pain (n=2); small-bowel obstruction (n=1); awaiting bowel function because of small-bowel injury (n=1); postoperative bleeding requiring reoperation (n=1); acute kidney injury (n=1); femoral neuropathy (n=1); migraine (n=1); postoperative anemia (n=1); correction of electrolyte abnormalities (n=1); thrombocytopenia (n=1); orthostatic hypotension (n=1); and central line thrombosis and infection (n=1). One patient was in the ICU preoperatively for severe anemia and blood transfusion. Only 1 patient had no documented indication for a prolonged hospital stay and 1 who requested an additional day for personal reasons.

Discussion

Same-day discharge after laparoscopic hysterectomy improves patient satisfaction and curbs procedural cost.^6,7 A prolonged LOS (>1 day) after RH was found in nearly 20% of the patients in the current study, which was higher than expected; thus, the current authors think that significant effort should be made to understand risk factors for increased LOS after RH. Subsequently, targeted interventions are needed to decrease LOS after RH. In the patient population analyzed for this study, the most commonly recorded reason in the medical records for having LOS>1 day was pain control.

Uterine weight was previously shown to be associated with increased operating time^8,9 and higher EBL⁸ in laparoscopic hysterectomy, without a significant association with LOS.^8,9 One study on RH in patients with large uteri from five different practices had similar findings with longer duration of surgery and greater EBL in patients with uteri weighing ≥500 g, without a significant difference in LOS.¹⁰ The current authors wanted to study this association of uterine weight and length of hospital stay further within a sample of RH patients from the same practice group and without limiting it to large uteri.

In addition, strategies that shrink uterine size preoperatively (such as gonadotropin-releasing hormone [GnRH] analogues) could decrease procedure duration and EBL,¹¹ and it would be helpful to know if they can decrease LOS or improve other postoperative outcomes. What is perhaps more important is whether women will tolerate the side-effects of GnRH analogues if their use is actually associated with a shortened stay, patient satisfaction, or improved clinical outcomes. In the current bivariate analysis as well as the unadjusted logistic regression model, uterine weight was associated with an increased LOS, but once adjustments were made for potential confounders, there did not seem to be an association between the two factors. Larger studies are needed to explore this relationship further and to detect any association if present.

All attempts should be made to decrease LOS after RH to decrease overall cost of the procedure and improve patient satisfaction. The current analysis has shown that procedure duration, EBL, baseline pain score, oophorectomy, and number of ports used are factors associated with extended LOS. Further studies are needed to explore these factors. In addition, building predictive indices that would estimate an individual patient's probability of an extended stay would be of great clinical utility, so that targeted interventions are applied preoperatively to optimize outcomes and decrease LOS, or to prepare patients adequately for their surgical and recovery experiences. For example, the number of ports used during surgery might be a potential modifiable factor that needs further investigation.

In the current study, some patients' characteristics were included that might predispose these patients to higher levels of postoperative pain or poor postoperative coping, such as an underlying chronic pain syndrome or history of depression or anxiety. It was found that baseline pain score was associated with increased LOS but history of depression or anxiety was not. This is an interesting area that needs further investigation. Interventions to optimize pain control prior to surgery are feasible and would be of benefit if the relationship between baseline pain and LOS holds to be true in further research.

Preoperative management of patient expectations of postoperative recovery, including time in the hospital, activity level, and pain level must also be addressed. When the reasons for extended LOS were investigated, it was found that almost 40% of the patients stayed in the hospital secondary to postoperative pain issues. Uncontrolled pain after ambulatory laparoscopic surgery is a major reason for delayed discharge or unanticipated readmission after surgical intervention.¹²

Often, patients who are counseled inadequately preoperatively regarding postoperative pain and activity limitations are more likely to request additional time in the hospital. Prior studies evaluating the “fast track principles” of recovery have suggested that effective preoperative counseling about postoperative pain, discharge timing, and general recovery expectations at the time of discharge result in decreased need for postoperative analgesia and improved overall recovery.¹³ In addition, optimizing inpatient pain management may also help maximize outcomes and reduce LOS.¹⁴ The current study we did not investigate postoperative pain management or pain scoring as this was not one of the predetermined outcomes; however, the finding of pain control as a reason for extended stay in the majority of patients was an important secondary finding.

The current study was limited by its retrospective design, the limited number of patients, and the fact that these results may not be generalizable to women in other parts of the United States. In addition, selection of patients was based on CPT codes, and the current authors could not exclude the possibility of miscoding or misclassification. Finally, it was not possible to identify the cases that were converted from laparoscopic hysterectomy to an open approach, as there is no CPT code for the conversion. Thus it was not possible to analyze the conversion cases separately.

Conclusions

In the patient population for this study, there was no clinically important association found between uterine size and extended LOS following RH. Further studies are needed to explore factors such as procedure duration, EBL, baseline pain score, oophorectomy, and number of ports used, any of which might be associated with extended LOS after RH. Pain management seemed to be the single greatest indication for increased LOS after RH. Indices predicting extended LOS after RH would be of significant clinical relevance to providers and patients.

Footnotes

Acknowledgments

The authors extend thanks to Andrew S. Bossick, MPH, in the department of public health sciences, in the Henry Ford Health System, Detroit, MI.

Disclosure Statement

No financial conflicts exist.

References

Wright

, Ananth

, Lewin

, et al. Robotically assisted vs laparoscopic hysterectomy among women with benign gynecologic disease. JAMA, 2013; 309:689.

van Dam

, Hauspy

, Verkinderen

, et al. Are costs of robot-assisted surgery warranted for gynecological procedures?. Obstet Gynecol Int, 2011; 2011:973830.

Pasic

, Rizzo

, Fang

, Ross

, Moore

, Gunnarsson

. Comparing robot-assisted with conventional laparoscopic hysterectomy: Impact on cost and clinical outcomes. J Minim Invasive Gynecol, 2010; 17:730.

Liu

, Lu

, Wang

, Shi

, Song

, Clarke

. Robotic surgery for benign gynaecological disease. Cochrane Database Syst Rev, 2012; 2:CD008978.

Sarlos

, Kots

. Robotic versus laparoscopic hysterectomy: A review of recent comparative studies. Curr Opin Obstet Gynecol, 2011; 23:283.

Chou

, Rosen

, Cario

, et al. Home within 24 hours of laparoscopic hysterectomy. Aust N Z J Obstet Gynaecol, 1999; 39:234.

Thiel

, Kamencic

. Assessment of costs associated with outpatient total laparoscopic hysterectomy. J Obstet Gynaecol Can, 2006; 28:794.

O'Hanlan

, McCutcheon

. Laparoscopic hysterectomy: Impact of uterine size. J Minim Invasive Gynecol, 2011; 18:85.

Kondo

, Bourdel

, Marengo

, et al. Surgical outcomes of laparoscopic hysterectomy for enlarged uteri. J Minim Invasive Gynecol, 2011; 18:310.

10.

Payne

, Dauterive

, Pitter

, et al. Robotically assisted hysterectomy in patients with large uteri: Outcomes in five community practices. Obstet Gynecol, 2010; 115:535.

11.

Lethaby

, Vollenhoven

, Sowter

. Efficacy of pre-operative gonadotrophin hormone releasing analogues for women with uterine fibroids undergoing hysterectomy or myomectomy: A systematic review. BJOG, 2002; 109:1097.

12.

Lovatsis

, Jose

, Tufman

, Drutz

, Murphy

. Assessment of patient satisfaction with postoperative pain management after ambulatory gynaecologic laparoscopy. J Obstet Gynaecol Can, 2007; 29:664.

13.

Donohoe

, Nguyen

, Cook

, et al. Fast-track protocols in colorectal surgery. Surgeon, 2011; 9:95.

14.

Wodlin

, Nilsson

. The development of fast-track principles in gynecological surgery. Acta Obstet Gynecol Scand, 2013; 92:17.