Goal-Directed Fluid Therapy in Autologous Breast Reconstruction allows Less fluid volume and more vasopressor use without outcome compromise

Structured Abstract

Objective:

Aggressive or restricted perioperative fluid management has been shown to increase complications in patients undergoing microsurgery. Goal directed fluid therapy (GDFT) aims to administer fluid, vasoactive agents, and inotropes according to each patient’s hemodynamic indices. This study assesses GDFT impact on perioperative outcomes autologous breast reconstruction (ABR) patients. We hypothesize that GDFT will have overall lower fluid administration and equivocal outcomes to patients not on GDFT.

Methods:

A single center retrospective review was conducted on ABR patients from January 2010-April 2017. An enhanced recovery pathway (ERP) using GDFT was implemented in April 2015. With GDFT, patients were administered intraoperative fluids, and vasoactive agents according to hemodynamic indices. Patients prior to April 2015 were considered pre-ERP cohort. Primary outcomes included amount and rate of fluid delivery, urine output (UOP), vasopressor administration, major (i.e. flap failure) and minor (i.e. seroma) complications, and length of stay (LOS).

Results:

Overall, 777 patients underwent ABR (ERP: 312; pre-ERP: 465). ERP patients received significantly less total fluid volume (ERP median: 3750 mL [IQR: 3000–4500 mL]; pre-ERP median: 5000 mL [IQR 4000–6400 mL]; p<0.001), had lower UOP, were more likely to receive vasopressor agents (47% vs 35%; p<0.001), and had lower LOS (ERP: 4 days[4–5]; pre-ERP: 5 [4–6]; p<0.001) compared to pre-ERP patients. Complications did not differ between cohorts.

Conclusions:

GDFT, as part of an ERP, and prudent use of vasopressors were found to be safe and did not increase morbidity in ABR patients. GDFT provides individualized perioperative care to the ABR patient.

Introduction

Autologous breast reconstruction (ABR) is considered the gold standard for post mastectomy breast reconstruction.¹ ABR typically requires lengthy operating times, which can exceed greater than eight hours and lead to fluid, blood, heat and losses/imbalances.^2–4Adequate perioperative management of systemic blood pressure is fundamental to ensuring sufficient flap perfusion, which are predominately maintained by intravenous fluid administration and vasopressors.⁵ Management of intraoperative fluid administration for ABR patients surround the use of aggressive, restricted, or goal-directed fluid therapy (GDFT).^6–9

Several retrospective analyses^{6, 8, 10–12} have suggested detrimental effects of overzealous intravenous fluid administration defined as exceeding 5.4 ml/kg/h¹³ was associated with flap complications in microsurgery. For example, Clark et al, reported that crystalloid volume greater than 130 mL/kg/ 24 h was an independent predictor for medical complications (OR 3.24) in head and neck microsurgery patients.¹³. Conversely, inadequate volume replacement in ABR can potentially lead to under-resuscitation and lead to decrease flap perfusion which can increase the risk of postoperative thrombosis.¹⁴ In a review of over 350 microsurgical breast reconstruction cases, 54 (20.8%) of patients had complications that in multivariate analysis suggested that the extremes of crystalloid infusion rate significantly predicted postoperative complications. The authors thus recommended to limit volume replacement to 6 ml/kg/h during the 24-h postoperative period.⁸

In order to standardize anesthetic management and prevent associated complications, such as decrease in flap perfusion leading to venous and arterial thrombosis, and medical complications associated with extremes of fluid management, this institution implemented an enhanced recovery pathway (ERP) protocol in April 2015.¹⁵ The ERP protocol includes GDFT, managing patients’ hemodynamic status by appropriately administering fluid, vasoactive agents, and inotropes according to cardiac output values and other hemodynamic indices (e.g. urine output [UOP])^16–19. A meta-analysis including 4,805 patients undergoing major abdominal surgeries perioperative GDFT was associated with a significant reduction in short-term mortality (risk ratio [RR] 0.75), long-term mortality (RR 0.80), and overall complication rates (RR 0.76) compared to the control group.¹⁶ Although, studies have shown perioperative vasopressor administration has an insignificant impact on ABR outcomes (i.e. flap failure),^{14, 20–22} microsurgeons remain reluctant to use perioperative vasopressor administration due to concern surrounding flap compromise¹⁶ and as the evidence for ABR patients remains limited.

The purpose of this study is to assess modifiable anesthetic management in patients undergoing ABR. The primary endpoint was assessing the impact of this institution’s transition to GDFT, including vasopressor administration, on 30-day complications, and overall resuscitation outcomes. We hypothesized that patients being managed by GDFT and prudent vasopressor administration as part of ERP protocol will have overall lower fluid administration and equivocal outcomes to pre-ERP cohort.

Methods

Study Design

Following approval from the Institutional Review Board at Memorial Sloan Kettering Cancer Center, an academic, National Cancer Institute-designated Comprehensive Cancer Center, we performed a retrospective cohort study to examine all ABR surgeries from January 2010 to April 2017. Inclusion criteria included patients 18 years and older who underwent immediate or delayed breast reconstruction with autologous tissue transfer using free transverse rectus abdominis myocutaneous (TRAM), muscle-sparing TRAM, deep inferior epigastric perforator (DIEP), superficial inferior epigastric artery perforator (SIEA), or pedicled TRAM (pTRAM), as well as patients who underwent therapeutic oncologic and/or prophylactic mastectomy. Exclusion criteria included patients who underwent simultaneous procedures, such as oophorectomy, and implant-based breast reconstruction, and patients younger than 18 years of age.

Perioperative Fluid Management Paradigm Shift

An ERP protocol for ABR patients was adopted at this institution in April 2015, which has been followed for all patients undergoing surgery since April 2015. Perioperative fluids were administered according to the this center’s GDFT regimen, as previously described.¹⁵ Goal-directed fluid therapy describes the use of a patient’s cardiac output values and other hemodynamic values to guide perioperative intravenous fluid and ionotropic therapy. Patients who received surgery prior to April 2015 were considered pre-ERP and did not receive GDFT. All patients who were part of the pre-ERP cohort had invasive arterial blood pressure monitoring, unless there was an absolute contraindication. Patients who were part of the ERP cohort underwent non-invasive monitoring, unless individual patient factors warranted invasive monitoring. Goal-directed fluid administration was guided by cardiac output values such as mean arterial pressure (MAP), heart rate, stroke volume, systemic vascular resistance and other hemodynamic indices including UOP and end tidal CO2. Mean arterial pressure (MAP) was consistently maintained between 65 and 85 mmHg.

Data Collection

A detailed review of electronic medical records was conducted. Variables recorded for each patient included age, body mass index (BMI), history of smoking, American Society of Anesthesiology (ASA) physical status classification, preoperative MAP, and hematocrit and creatine lab values. Data collected on comorbidities included asthma, diabetes, chronic obstructive pulmonary disease (COPD), hypertension, sleep apnea, hyperlipidemia, cardiomyopathy, peripheral artery disease, coronary artery disease, arrhythmia, history of pain, and history of nausea and vomiting.

Preoperative and postoperative data on breast irradiation, neoadjuvant/adjuvant chemotherapy, hormone therapy, free-flap type, and reconstructive timing were recorded. Reconstructive timing was defined as either immediate (at the time of mastectomy) or delayed. Intraoperative data included duration of surgery, intraoperative fluid management, estimated blood loss (EBL), urine output (UOP), type and amount of vasopressor administration, and hemodynamic parameters such as blood pressure and heart rate readings.

Outcomes of Interest

The primary outcome of interest was 30-day postoperative complications (major and minor). Major complications were defined as those requiring an unplanned return to the operating room (OR) or hospital readmission. Examples of major complications included breast cellulitis requiring IV antibiotics, breast hematoma requiring OR drainage, total flap loss, and venous or arterial thrombosis. Complications that were characterized as minor included breast cellulitis not requiring IV antibiotics, hematoma, seroma, mastectomy skin flap necrosis, partial flap loss, breast wound-related complications, donor-site complications, pneumonia, postoperative blood transfusion, and pulmonary embolism. The secondary outcome of interest was hospital length of stay (LOS), defined as the number of days from initial admission to hospital discharge and 30-day readmission to the hospital rates.

Statistical Analysis

The primary analysis compared patients who were administered GDFT with patients who underwent standard fluid resuscitation without a standardized regimen. Secondarily, patients were examined based on the degree of resuscitation using the rate of fluid administration and UOP as real time indicators of invasive monitoring. Rate of UOP and rate of total fluid administered were categorized as below normal (lowest 16^th percentile), normal (17^th–84^th percentile), and above normal (top 16^th percentile), which represented one standard deviation above and below the mean total fluid administered. The Kruskal-Wallis test was used for continuous variables, and Fisher’s exact test was used for categorical variables in comparisons between fluid and urine levels. SAS version 9.4 (SAS Institute Inc., Cary, NC) was used for all analysis. All tests were two-sided and p < 0.05 was considered significant.

Results

ERP vs Pre-ERP Cohorts

Overall, 777 patients who underwent ABR from January 2010 to April 2017 were included in the study; 312 patients were ERP (GDFT cohort) and 465 patients were pre-ERP (pre-GDFT cohort). The median age for the overall cohort was 51 years (ERP: 50 [IQR: 44–55]; pre-ERP: 51 [IQR: 44–58]; p=0.124). As shown in Table 1, most patient demographics and preoperative characteristics were similar in both cohorts. There was no significant difference between the ERP and pre-ERP groups in the proportion of delayed and immediate ABR and in the laterality of reconstruction. Type of ABR significantly differed between the two groups (p<0.001),.

Table 1.

Demographic and surgical case characteristics pre-ERP (pre-GDFT) and during ERP (GDFT)

	Total		Pre-ERP (pre-GDFT)		ERP (GDFT)		p-value*
	(n=777)		(n=465)		(n=312)
Age (years)	51	(44–56)	50	(44–55)	51	(44–58)	0.124
BMI (kg/m2)	29	(25.90–32.70)	29.3	(26.00–32.80)	28.8	(25.65–32.65)	0.209
Diabetes	61	7.90%	41	8.8%	20	6.4%	0.276
Hypertension	168	21.60%	106	22.8%	62	19.9%	0.374
Coronary Artery Disease	3	0.40%	0	0.0%	3	1.0%	0.064
Peripheral Artery Disease	2	0.30%	2	0.4%	0	0.0%	0.519
Hyperlipidemia	161	20.70%	97	20.9%	64	20.5%	0.928
Active or Former Smoker	269	34.60%	163	35.10%	106	34%	0.818
Asthma	76	9.80%	46	9.90%	30	9.60%	>0.95
COPD	5	0.60%	4	0.90%	1	0.30%	0.653
Obstructive Sleep Apnea	68	8.80%	44	9.5%	24	7.7%	0.438
History of Arrhythmia	20	2.60%	9	1.9%	11	3.5%	0.175
History of Pain	187	24.10%	102	21.9%	85	27.2%	0.104
History of Nausea and Vomiting	246	31.70%	160	34.4%	86	27.6%	0.049
History of Chemotherapy	348	44.80%	216	46.5%	132	42.3%	0.27
History of Radiation	317	40.80%	195	41.9%	122	39.1%	0.457
History of Hormone Therapy	237	30.50%	147	31.6%	90	28.8%	0.428
ASA Physical Status							0.608
1 or 2	413	53.20%	251	54.0%	162	51.9%
3 or 4	364	46.80%	214	46.0%	150	48.1%
Procedure Timing							0.299
Immediate	458	58.90%	267	57.4%	191	61.2%
Delayed	319	41.10%	198	42.6%	121	38.8%
Laterality							0.103
Bilateral	330	42.50%	186	40.0%	144	46.2%
Unilateral	447	57.50%	279	60.0%	168	53.8%
Flap Type							<0.001
pTRAM flap	123	15.80%	102	21.9%	21	6.7%
msTRAM flap	210	27%	146	31.4%	64	20.5%
DIEP flap	392	50.50%	193	41.5%	199	63.8%
Other	52	6.70%	24	5.2%	28	9.0%
Preoperative MAP	91.3	(84.0–98.3)	91	(83.3 – 98.0)	92.2	(85.0–98.7)	0.098
Preoperative Hematocrit (%)	38.5	(36.6–40.3)	38.2	(36.4–40.1)	39	(36.9–40.9)	0.002
Preoperative Creatinine (mg/dL)	0.8	(0.70–0.80)	0.7	(0.7 – 0.9)	0.8	(0.7 – 0.8)	0.07

Continuous data are expressed as median (range). Categorical data are expressed as n (percent).

^*p value calculated with Kruskal-Wallis (continuous) or Fisher’s exact test (categorical).

ASA indicates American Society of Anesthesiologists; BMI, body mass index; COPD, chronic obstructive pulmonary disease; DIEP, deep inferior epigastric perforator; ERAS, enhanced recovery after surgery; GDFT, goal-directed fluid therapy; MAP, mean arterial pressure; msTRAM = muscle-sparing transverse rectus abdominis myocutaneous. pTRAM = pedicled transverse rectus abdominis myocutaneous.

The median anesthesia length was significantly shorter in the ERP group than in the pre-ERP group (ERP: 9.97 hours [IQR: 7.98–11.60]; pre-ERP: 10.40 hours [IQR: 8.40–12.10]; p=0.021; Table 2]. Overall, ERP patients received significantly less median total fluid volume (ERP: 3750 mL [3000–4500 mL]; pre-ERP: 5000 mL [4000–6400 mL]; p<0.001) at a significantly lower rate (p < 0.001). Total fluid balance, UOP, and EBL were all significantly lower in the ERP group, compared to the pre-ERP group (p< 0.001). Median intraoperative MAP was significantly lower among ERP patients (ERP: 69 mmHg [IQR: 66–72]; pre-ERP: 71 mmHg [67–75]; p<0.001). The overall proportion of patients with vasopressor agents administered (p<0.001), including the specific administration of phenylephrine (p=0.039) and ephedrine (p<0.001), was significantly higher in the ERP group, compared to pre-ERP. Postoperative major and/or minor complications (p=0.406) and 30-day major complications (p=0.467) did not differ between cohorts nor did the proportion of patients who returned to the OR (p=0.809) or were readmitted (p=0.884) within 30 days. Median LOS was significantly lower among ERP patients (ERP: 4 days [4–5]; pre-ERP: 5 [4–6]; p<0.001).

Table 2.

Perioperative outcomes pre-ERP (pre-GDFT) and during ERP (GDFT).

	Total		Pre-ERP (pre-GDFT)		ERP (GDFT)		p-value*
	(n=777)		(n=465)		(n=312)
Intraoperative Outcomes
Anesthesia Length (hours)	10.2	8.27– 12.00	10.4	8.40 – 12.1	9.97	7.98 – 11.6	0.021
Total Fluids (mL)	4500	(3500–5700)	5000	(4000–6400)	3750	(3000–4500)	<0.001
Total Fluid Administration Rate (mL/hr)	448	(371–530)	500	(429–575)	376	(321–442)	<0.001
Total Crystalloid (mL)	4000	(3300–5200)	4700	(3900–6000)	3500	(2750–4000)	<0.001
Crystalloid Infusion Rate (mL/hr)	416	(344–492)	458	(402–534)	348	(297–400)	<0.001
Total Colloid (mL)	250	(0.00–500)	250	(0–750)	250	(0–500)	0.016
Colloid Infusion Rate (mL/hr)	26.2	(0.00–59.4)	29.8	(0–64.7)	24.2	(0–52.2)	0.07
Urine Output (mL)	650	(450–930)	680	(500–1000)	600	(403–850)	<0.001
Total Urine Output Rate (mL/hr)	63.3	(47.7–89.5)	65.9	(50.5–95.5)	59.1	(44.8–81.7)	<0.001
Estimated Blood Loss (mL)	150	(100.00–200.00)	150	(100–250)	100	(100–200)	<0.001
Total Fluid Balance (mL)	3770	(2830–4850)	4320	(3450–5400)	2965	(2418–3843)	<0.001
Vasopressors Administered	310	39.90%	163	35.10%	147	47.10%	<0.001
Phenylephrine Administered	146	18.80%	76	16.30%	70	22.40%	0.039
Ephedrine Administered	235	30.20%	118	25.40%	117	37.50%	<0.001
Intraoperative Heart Rate	74	(67.0–80.0)	74	(68.0–80.0)	73	(66.0–79.0)	0.03
Intraoperative MAP	70	(67.0–74.0)	71	(67.0–75.0)	69	(66.0–72.0)	<0.001
Intraoperative Thromboses	6	0.77%	4	0.86%	2	0.64%	>0.95
Postoperative Outcomes
Postoperative MAP	85	(77.7–93.3)	85.7	(79.0–95.3)	83	(76.0–91.3)	0.006
Postoperative Hematocrit (%)	32	(29.1–34.4)	32.3	(29.4–34.7)	31.7	(28.8–34.0)	0.06
Postoperative Creatinine (mg/dL)	0.7	(0.60–0.80)	0.7	(0.60–0.80)	0.7	(0.60–0.80)	0.003
Postoperative Thromboses	26	3.35%	18	3.87%	8	2.56%	0.68
Length of Stay (days)	5	(4.00–6.00)	5	(4–6)	4	(4–5)	<0.001
Minor and/or Major Complication	291	37.50%	180	38.70%	111	35.60%	0.406
30-Day Major Complication	111	14.30%	70	15.10%	41	13.10%	0.467
30-Day Return to Operating Room	79	10.20%	46	9.90%	33	10.60%	0.809
30-Day Readmission to Hospital	52	6.70%	32	6.90%	20	6.40%	0.884

Continuous data are expressed as median (IQR). Categorical data are expressed as n (percent).

^*p value was calculated with Kruskal-Wallis (continuous) or Fisher’s exact test (categorical).

ERP indicates enhanced reovery pathway; GDFT, goal-directed fluid therapy; MAP, mean arterial pressure.

Intraoperative Rate of Total Fluid Administered Cohorts

Analysis of intraoperative total fluid delivery by rate of administration in both the GDFT and pre-GDFT groups showed that patients in the below normal fluid administration cohort (lowest 16^th percentile) received fluid at a rate of less than 314 mL/hr, and patients in the above normal cohort (top 16^th percentile) received fluid at a rate of greater than 577 mL/hr (Table 3). The three cohorts, including patients in the normal fluid administration group (17^th–84^th percentile), had similar baseline characteristics and past medical history, although the above normal cohort had a higher proportion of patients with obstructive sleep apnea (p=0.039).

Table 3.

Patient and case characteristics by total fluid administration rate (ml/hr).

	Rate of total fluid administration (ml/hr)
	Below Normal Total Fluid Admin Rate		Normal Total Fluid Admin Rate		Above Normal Total Fluid Admin Rate		p-value*
	(<314 mL/hr)		(314–577 mL/hr)		(>577 mL/hr)
	(n=124)		(n=530)		(n=123)
Age (years)	49	(43–57)	51	(44–56)	51	(45–55)	0.679
BMI (kg/m2)	29.5	(26.1–33.0)	28.9	(25.7–32.4)	29.5	(26.0–33.9)	0.172
Diabetes	9	7.30%	45	8.50%	7	5.70%	0.612
Hypertension	20	16.10%	116	21.90%	32	26%	0.162
Coronary Artery Disease	0	0%	2	0.40%	1	0.80%	0.46
Peripheral Artery Disease	0	0%	1	0.20%	1	0.80%	0.317
Hyperlipidemia	27	21.80%	108	20.40%	26	21.10%	0.91
Active or Former Smoker	48	38.70%	184	34.70%	37	30.10%	0.362
Asthma	12	9.70%	49	9.20%	15	12.20%	0.581
COPD	0	0%	4	0.80%	1	0.80%	0.827
Obstructive Sleep Apnea	5	4%	47	8.90%	16	13%	0.039
History of Arrhythmia	2	1.60%	13	2.50%	5	4.10%	0.481
History of Pain	27	21.80%	134	25.30%	26	21.10%	0.528
History of Nausea and Vomiting	40	32.30%	168	31.70%	38	30.90%	>0.95
History of Chemotherapy	47	37.90%	243	45.80%	58	47.20%	0.236
History of Radiation	44	35.50%	225	42.50%	48	39%	0.339
History of Hormone Therapy	32	25.80%	167	31.50%	38	30.90%	0.463
ASA physical status							0.482
1 or 2	71	57.30%	281	53%	61	49.60%
3 or 4	53	42.70%	249	47%	62	50.40%
Procedure Timing							0.557
Immediate	77	62.10%	313	59.10%	68	55.30%
Delayed	47	37.90%	217	40.90%	55	44.70%
Laterality							0.451
Bilateral	55	44.40%	229	43.20%	46	37.40%
Unilateral	69	55.60%	301	56.80%	77	62.60%
Flap Type							0.053
pTRAM flap	17	13.70%	81	15.30%	25	20.30%
msTRAM flap	23	18.50%	159	30%	28	22.80%
DIEP flap	76	61.30%	252	47.50%	64	52%
Other	8	6.50%	38	7.20%	6	4.90%
Preoperative MAP	90.5	(83.3–97.3)	91.7	(85.0–99.0)	90.7	(81.3–97.3)	0.199
Preoperative Hematocrit (%)	38.9	(36.6–41.0)	38.5	(36.6–40.3)	38.2	(36.2–39.5)	0.124
Preoperative Creatinine (mg/dL)	0.8	(0.70–0.80)	0.8	(0.70–0.80)	0.7	(0.60–0.80)	0.022

Continuous data are expressed as median (range). Categorical data, expressed as n (percents). BMI = body mass index. COPD = chronic obstructive pulmonary disease. ASA = American Society of Anesthesiologists physical status. pTRAM = pedicled transverse rectus abdominis myocutaneous. msTRAM = muscle-sparing transverse rectus abdominis myocutaneous. DIEP = deep inferior epigastric perforator.

^*p value calculated with Kruskal Wallis (continuous) or Fisher’s Exact test (categorical)

Regarding intraoperative characteristics, the median anesthesia length significantly differed among the three groups, with the normal cohort having the longest possible duration (p=0.048; Table 4). As the rate of fluid administration increased (moving from below normal to normal to above normal fluid rates), median postoperative creatinine and median postoperative hematocrit decreased significantly (p < 0.02). Increased rates of fluid administration also resulted in increased total fluid volume, total crystalloid volume, total crystalloid infusion rate, total colloid volume, colloid infusion rate, UOP, EBL, and total fluid balance; these increases were significant (p <0.001). However, postoperative complications, 30-day return to the OR, and 30-day readmission outcomes did not differ significantly, based on fluid administration rate.

Table 4.

Perioperative outcomes by total fluid administration rate (ml/hr).

	Rate of total fluid administration (ml/hr)
	Below Normal Total Fluid Admin Rate		Normal Total Fluid Admin Rate		Above Normal Total Fluid Admin Rate		p-value*
	(<314 ml/hr)		(341–577 ml/hr)		(>623 ml/hr)
	(n=124)		(n=530)		(n=123)
Anesthesia Length (hours)	10.1	(8.3–11.9)	10.1	(8.47–12.1)	9.8	(7.7–11.5)	0.048
Total Fluids (ml)	3000	(2500–3775)	4500	(3700–5500)	6150	(5000–7500)	<.001
Total Crystalloid (ml)	2900	(2400–3500)	4000	(3500–5000)	5800	(4700–6800)	<.001
Crystalloid Infusion Rate (ml/hr)	287	(261–308)	416	(367–461)	595	(559–647)	<.001
Total Colloid (ml)	0	(0.00–250)	250	(0.00–500)	750	(250–1000)	<.001
Colloid Infusion Rate (ml/hr)	0	(0.00–24.8)	27	(0.00–57.6)	67.8	(25.6–94.6)	<.001
Urine Output (ml)	550	(400–748)	660	(460–930)	710	(495–1310)	<.001
Estimated Blood Loss (ml)	100	(75.00–150.00)	150	(100.00–200.00)	200	(100.00–300.00)	<.001
Total Fluid Balance (ml)	2410	(1960–2858)	3775	(3000–4600)	5450	(4380–6325)	<.001
Vasopressors Administered	47	37.90%	211	39.80%	52	42.30%	0.787
Phenylephrine Administered	23	18.50%	93	17.50%	30	24.40%	0.216
Ephedrine Administered	36	29%	164	30.90%	35	28.50%	0.844
Median Intraoperative Heart Rate	74	(66.0–80.0)	74	(68.0–80.0)	74	(68.0–79.0)	0.724
Median Intraoperative MAP	70	(67.0–72.0)	70	(66.0–74.0)	70	(67.0–75.0)	0.261
Postoperative Outcomes
Postoperative MAP	85.3	(78.0–93.3)	84.7	(77.3–93.3)	85	(78.0–96.7)	0.703
Postoperative Hematocrit (%)	32.8	(30.2–35.9)	32	(29.1–34.4)	31.7	(29.0–33.5)	0.011
Postoperative Creatinine (mg/dL)	0.7	(0.60–0.80)	0.7	(0.60–0.80)	0.7	(0.60–0.70)	<.001
Length of Stay (days)	4	(4.00–5.00)	5	(4.00–6.00)	5	(4.00–6.00)	0.013
Minor and/or Major Complication	45	36.30%	207	39.10%	39	31.70%	0.307
30-Day Major Complication	17	13.70%	77	14.50%	17	13.80%	>0.95
30-Day Return to Operating Room	11	8.90%	57	10.80%	11	8.90%	0.788
30-Day Readmission to Hospital	10	8.10%	34	6.40%	8	6.50%	0.799

Continuous data are expressed as median (range). Categorical data, expressed as n (percents). MAP = mean arterial pressure.

^*p value calculated with Kruskal Wallis (continuous) or Fisher’s Exact test (categorical)

MAP indicates mean arterial pressure.

UOP

UOP, used as a proxy for fluid resuscitation, was categorized as below normal (< 43 mL/hr; n=120), normal (43–105 mL/hr; n=534), and above normal (> 105 mL/hr; n=123). For major complication (p=0.396), any complication (p=0.739), OR return (p=0.126), and hospital readmission (p=0.136), there were no significant differences between the three UOP cohorts (Table 5).

Table 5.

30-day postoperative outcomes by total urine output rate (mL/hr).

	Total Urine Output Rate
	Below Normal		Normal		Above Normal		p-value*
	(<43 mL/hr)		(43–105 mL/hr)		(>105 mL/hr)
	(n=120)		(n=534)		(n=123)
Major Complication	18	15%	71	13.30%	22	17.90%	0.396
Any Complication	44	36.70%	197	36.90%	50	40.70%	0.739
Return to OR	8	6.70%	53	9.90%	18	14.60%	0.126
Hospital Readmit	13	10.80%	33	6.20%	6	4.90%	0.136

Continuous data are expressed as median (range). Categorical data are expressed as n (percent).

^*p value calculated with Kruskal-Wallis (continuous) or Fisher’s exact test (categorical).

Discussion

This study of nearly 800 patients who underwent ABR over a seven-year period assessed the impact of fluid management and vasopressor use as part of the implementation of an ERP protocol at this institution. The ERP cohort received more vasopressors and less total intraoperative fluids at a lower rate than the pre-ERP group, and there were no differences in 30-day major or minor complications or readmission rates between the cohorts.

The objective of fluid delivery during surgery is to optimize cardiac performance as well as maintain and maximize tissue-oxygen delivery to promote optimal flap perfusion.⁹ Multiple studies have shown that clinical outcomes can be adversely affected by inadequate or extreme perioperative fluid administration.^{8 6, 7} Excessive fluid administration was the traditional approach for perioperative fluid management for ABR procedures,²³ based on the rationale that mild hypervolemia diminishes sympathetic activity and consequently dilates and promotes microvascular blood flow across the anastomotic site.^{23, 24} This approach has fallen out of favor as studies have shown that excess amounts of fluid administration can trigger inflammatory factors, increase clotting rates, promote excessive interstitial edema, and increase deleterious flap outcomes (i.e. thrombotic complications and wound breakdown).^{6, 8, 10, 11} For example, in patients undergoing head and neck free-flap reconstruction for cancer, Clarke et al. found that aggressive fluid management exceeding 5.4 ml/kg/h s was associated with flap complications, including arterial and venous thrombosis as well as partial and total flap loss.¹³ Similarly, Zhong et al. identified crystalloid infusion rate as a predictor of postoperative morbidity in microsurgical breast reconstruction and recommended a volume replacement limit of 6 ml/kg/h during the 24-h postoperative period.⁸ Conversely, inadequate volume replacement in these patients can also potentially lead to poor flow in the flap and leave patients at increased risk of delayed postoperative thrombosis.¹⁴

This center’s implementation of the ERP protocol in April 2015 has allowed for the tailoring of fluid management by using GDFT in addition to the safe use of inotropes. Historically, during the early study period, invasive monitoring, such as arterial monitoring (pulse pressure and stroke volume), was commonly used to guide and direct the administration of fluids with the GDFT technique. However, with the implementation of the ERP protocol, non-invasive techniques such as finger cuffs to measure dynamic and flow-based perfusion parameters (stroke volume and stroke volume variation) have been employed.^25–27 Risks of complications are minimal with this non-invasive monitoring method, so it is now much preferred over invasive monitoring.^27–29

The most significant finding in this study was that GDFT patients did not have higher postoperative complications or morbidity than pre-GDFT patients, despite having less volume resuscitation and more vasopressors administered. The use of vasopressors for the management of hypotension in microsurgery is controversial.³⁰ Multiple studies have advocated not using vasoactive agents in this setting; however, these studies are based on animal trials rather than on evidence-based medicine.^{31, 32} A recently published prospective randomized clinical trial on 44 patients undergoing DIEP flaps were randomized to either fluid restrictive vasopressor dominated or vasopressor restrictive and liberal fluid administration support to assess total flap perfusion detected via indocyanine green fluorescence (IGF) imaging. The authors found that there was no significant difference in total flap microperfusion between cohorts when assessed on IGF. Furthermore, consistent with current literature, there was a rise in complications with intraoperative fluid over-resuscitation.³³ The study adds to the growing literature, that neither the administration nor type of vasopressor (phenylephrine and ephedrine) impacted outcomes. Consistent with other clinical studies, the intraoperative use of vasopressors in this study did not affect perioperative surgical outcomes nor lead to higher complication rates.^{14, 20–22, 34}

It is important to note that the implementation of this institution’s ERP protocol significantly decreased LOS. Accordingly, ERP patients were counseled preoperatively on a shorter LOS, and the timeline for postoperative milestones (i.e., advancement of diet, ambulation, etc.) were adjusted. This confounded the ability for us to assess if the perioperative anesthetic management itself impacted postoperative LOS, and likely did not play a significant role in changing this endpoint.

Cosmetic outcomes and long-term patient-reported outcomes in microsurgery have been continually focused on for improving practice and advancing the field.^{7, 15, 35} Perioperative care is an area that has gained traction in recent years in an effort to further individualize and tailor patient care. The appropriate intraoperative administration of fluids is a crucial aspect of individualized plans for fluid and hemodynamic management. The additional benefit of GDFT should be determined based on surgical and patient risk factors. GDFT should not be used in isolation, instead, perioperative hemodynamic trends and the fluid priorities of the patient should always be considered. This study’s results may promote a paradigm shift to GDFT, which has demonstrated reduced amounts and rates of fluid received by patients undergoing ABR.

Although this study highlights essential aspects of anesthetic management for patients undergoing ABR, it is not without its limitations. The retrospective nature of this study limits the extent to which conclusions may be drawn, in addition to the inherent biases in data collection from a single site. Since the anesthesia team was not constant throughout the study, there may be small differences in practice patterns. The type of autologous procedure changed during the study period in that after ERP implementation in April 2015, nearly all flaps were free tissue transfer, which is important because, in theory, one would be more concerned with vasopressor delivery for free tissue transfers. The operative time in the ERP group was significantly shorter than that in the pre-ERP group. This finding is likely a result of the recent trend at our institution to perform microsurgical cases with two microsurgeons to reduce the patient’s time under general anesthesia. Another consideration is the increase in surgeon experience with perforator flaps. In the ERP cohort, comprised almost completely of free tissue transfers, vasopressor usage rates were higher, yet complications were similar. To further examine intraoperative anesthetic management as it relates to fluid delivery, vasopressor administration, and hemodynamic adjustment in the context of ABR, a prospective trial would need to be developed. Ideally, this trial would span multiple centers to adjust for biases among techniques at individual institutions. Additionally, this study may not be powered to determine small differences in rare complications (i.e. thrombotic complications), and the results should be viewed as such.

Conclusions

GDFT, as part of an ERP protocol, and prudent use of vasopressors were found to be safe and did not increase postoperative complications or morbidity in ABR patients. GDFT should be used to provide individualized intraoperative care to the ABR patient.