Comparison of Clinical Outcomes Following Head and Neck Surgery Among Patients Who Contract to Abstain From Alcohol vs Patients Who Abuse Alcohol

Key Points

Question

Does preoperative alcohol abstinence contracting improve postoperative outcomes?

Findings

In this case-control study, a clinical protocol was developed in which misusers of alcohol were asked to sign a contract to be abstinent of alcohol prior to surgery. This protocol was associated with a reduction in hospital stay, fewer wound complications, decreased withdrawal, and reduced time lapse to adjuvant radiation therapy.

Meaning

Empowering patients with abstinence contracting has potential to improve quality metrics in head and neck surgery.

Abstract

Importance

Alcohol abuse is highly prevalent in the population of patients with head and neck cancer, and active abuse at the time of surgery results in poor postoperative outcomes.

Objective

To determine the association of alcohol abstinence with postoperative outcome in alcohol misusers undergoing major surgical procedures of the head and neck.

Design, Setting, and Participants

As a case-control study, from July 2013 to August 2015, outcomes were collected in patients at an academic tertiary referral center requiring free flap reconstruction who had undergone an abstinence contract (n = 15) and compared with those who were abusers of alcohol prior to the inception of the protocol (n = 30).

Interventions

Our institution developed a clinical protocol in which patients who are misusers of alcohol are educated and voluntarily asked to sign an alcohol abstinence contract agreeing to be abstinent of alcohol a minimum of 7 days prior to surgery.

Main Outcomes and Measures

Rate of alcohol withdrawal, length of stay, return to operating room, and readmission within 30 days were determined from medical record review.

Results

In the contracted group (15 patients), the mean age was 62 years, and 12 were male; in the control group (30 patients), the mean age was 58 years, and 26 were male. In both cohorts, no statistically significant difference was found between initial staging, surgery performed, and preoperative morbid conditions. The alcohol abstinence contracted group had an average of 14.8 days of abstinence prior to surgery. Abstinence was achieved at home for 11 patients, in a detoxification facility for 2 patients, and with preoperative admission for 2 patients. In direct comparisons between the groups, the rates of alcohol withdrawal (63% vs 0%; difference, 0.63; 95% CI, 0.33-0.85), delirium (73% vs 0%; difference, 0.73; 95% CI, 0.45-0.92), cellulitis (43% vs 7%; difference, 0.37; 95% CI, 0.04-0.64), and wound dehiscence (67% vs 13%; difference, 0.53; 95% CI, 0.22-0.79) were higher in the noncontracted group vs the contracted group. Furthermore, hospital stay (median 13 days vs 9 days; difference, 5 days; 95% CI, 3-7 days) and time lapse to starting adjuvant radiation therapy (median, 60.0 days vs 42.5 days; difference, 15 days; 95% CI, 5.0-25 days) were statistically and clinically significantly longer in the noncontracted group vs the contracted group.

Conclusions and Relevance

An alcohol abstinence program for surgically treated patients is safe and seems to be effective in reducing morbidity and improve outcomes.

This case-control study evaluates the association of alcohol abstinence with postoperative outcome in alcohol misusers undergoing major surgical procedures of the head and neck by using an alcohol abstinence contract.

Introduction

Recent studies have shown that patients who drink at least 60 g of alcohol (4-5 standard US drinks; 1 standard US drink = 14 g of alcohol) per day have a 3-fold increase in postoperative complications. Alcohol misusers have been shown to have increased hospital stay, increased need for second operations, and more postoperative myocardial events compared with nonusers. From a mechanistic standpoint, alcohol leads to these adverse events secondary to alcohol-induced cardiomyopathy, destruction of normal sleep architecture (resulting in hypoxemia), immune suppression, coagulopathy, delayed soft-tissue healing, and withdrawal. Postoperative delirium alone leads to higher rates of complications, higher mortality, and increased hospital stays (15 vs 7 days). Alcohol withdrawal generally presents as delirium, agitation, delirium tremens, tachycardia, and fever. A withdrawing patient may place their surgical wounds at risk of disruption from their agitation. The burden of postoperative complications, both economically and medically, is most damaging when reconstructing head and neck defects with microvascular free tissue transfer. Furthermore, the prolonged hospital admission for free flap patients precipitates withdrawal which typically occurs 24 to 72 hours after the last drink. To mitigate alcohol withdrawal, many institutions have adopted a standardized protocol for postoperative administration of benzodiazepine therapy. Despite this effort, the rate of withdrawal remains unacceptably high, particularly in heavy alcohol misusers.

In general, the duration of abstinence to counteract the negative effect of alcohol ranges from 1 to 8 weeks, depending on the organ system. Alcohol abstinence for just 2 weeks has been shown to effectively recover platelet function and reduce bleeding times. The risk of withdrawal and delirium is known to diminish after 1 week of abstinence. The length of time for abstinence must be weighed with the risk of delaying surgery in the population of patients with cancer, but in high-volume centers, 30 days from initial clinic visit is acceptable and does not lead to untoward effects from an oncologic standpoint.

To our knowledge, no study has yet been done to evaluate the effect of alcohol abstinence in head and neck patients who misuse alcohol. This group deserves special attention for intervention because the prevalence of alcoholism among the head and neck patient is as high as 30% to 90%. We have an aggressive postoperative alcohol withdrawal protocol in place to minimize the effects of alcohol withdrawal. However, these protocols are reactionary and do not prevent severe withdrawal in heavy misusers. In response to this, our multidisciplinary team developed a preventative strategy with an alcohol cessation protocol for the heavy alcohol misusers to augment the current postoperative protocol. The objective of this study was to determine if there was any improvement in patient outcomes since the advent of our alcohol cessation protocol and to report our early findings.

Methods

Clinical Protocol

In July 2013, the Division of Head and Neck Surgery at our institution implemented a clinical protocol for alcohol misusers. This protocol was developed in conjunction with the departments for internal medicine, social work, nursing, and otolaryngology–head and neck surgery. Alcohol misusers are identified at the time of initial visit by in-office CAGE (Cutting down, Annoyance by criticism, Guilty feeling, and Eye-openers) questionnaires, and the protocol is applied at the discretion of the attending surgeon based on the physician-patient relationship, severity of the cancer, and need for surgical intervention. They are requested to voluntarily sign an alcohol abstinence contract (Box), which asks that they abstain from all alcohol consumption for a minimum of 7 days prior to surgery. If they agree, the protocol is implemented (Figure). To bolster compliance, contracts are signed in the presence of family, and patients are scheduled to visit with the staff from internal medicine department preoperatively. The nursing and social work departments are notified and call the patient periodically to confirm compliance. If patients are deemed to be too high risk to undergo abstinence at home or lapse and are not compliant with the contract, the option to be admitted to a detox facility or referred for inpatient admission is available. Patients who are treated with the home protocol are given a benzodiazepine prescription with instructions to take if signs or symptoms of withdrawal occur. In addition, patients are educated regarding concurrent pain medication and benzodiazepine use. Ultimately, if a patient refuses to sign the contract, surgery will not be withheld, but strong consideration for inpatient preoperative admission or admission to a detoxification facility 7 days prior to surgery is discussed with the patient and family by the physician, social workers, and nursing staff. We have never withheld surgery from our patients who are compliant with all other preoperative testing. All patients are assessed with the Clinical Institute Withdrawal Assessment for Alcohol (CIWA) scale postoperatively as described earlier.

Box. The Alcohol Abstinence Contract Given to Patients With a History of Alcohol Abuse.

• I understand that continuing to drink alcohol increases my risk for serious complication after surgery. Such complications include confusion with violent behavior causing me to hurt myself and others, infections such as wound infection, pneumonia and sepsis (blood poisoning), excessive bleeding, and death.

• I will NOT drink any alcohol starting today. This means not a near-beer, not a half a beer, not even a sip of alcohol of any kind. NONE. I understand that, if I drink any alcohol after today this may cause my surgery to be cancelled. I have discussed with my caregivers the steps I need to take to stop drinking alcohol and the resources available to help me stop drinking alcohol. I understand that I will be subject to random alcohol and drug testing at any point during my treatment and I agree to such testing.

• Sincerely,

• (Patient’s signature)

• _______________________________

Figure. Algorithm for Implementation of Abstinence Contract.

CAGE scale: Cutting down, Annoyance by criticism, Guilty feeling, and Eye-openers) questionnaire.

Data Collection and Statistical Analysis

After obtaining approval from the Ohio State University institutional review board, a retrospective medical record review was performed for patients requiring free flap reconstruction who had undergone a cessation protocol (from July 2013 to August 2015). Waiver of the informed consent process was obtained given the retrospective nature of our study. Patients were no compensated for their participation. Patients undergoing free flap reconstruction were chosen to minimize the variables because this is a relatively standardized group treated with the same postoperative free flap protocol (non–intensive care unit protocol), which has been previously published. A control group was created by identifying patients who were abusers of alcohol prior to the inception of the alcohol cessation protocol but treated with the CIWA protocol to minimize postoperative variability. Previous abusers of alcohol were found by searching our surgical logs from June 2007 to June 2013 for patients who reported consuming more than 7 US drinks per day because real-time patient interviews were not possible. The rationale for this number of drinks was based on institutional experience that patients reporting less than 7 drinks per day were successfully treated with the CIWA protocol alone and did not experience major complications. Those reporting 7 or more drinks per day were defined as the high-risk population. Medical records were reviewed for preoperative characteristics (staging, primary subsite, history of radiation therapy, smoking status), as well as preexisting comorbidities (renal disease, hepatic disease, pulmonary disease, vascular disease, diabetes, hypothyroidism, coronary artery disease, obesity, malnutrition). Postoperative complications (flap loss, myocardial infarction, cerebrovascular accident, alcohol withdrawal, delirium, fistula formation, surgical site infection, pneumonia, death, and cardiopulmonary arrest) were also tabulated. To determine the impact on level of service of our intervention, we also recorded length of stay, rate of return to operating room during admission and readmission to the hospital within 30 days. Length of stay was calculated from day of surgery until discharge. Finally, the time lapse to begin radiation therapy after date of surgery was determined. Primary outcomes include alcohol withdrawal, length of stay, return to operating room and readmission within 30 days. Secondary outcomes include the other postoperative complications listed herein and time lapse to radiation therapy. Comparisons between the noncontracted and contracted groups for age was made by 2-sample t test and the other continuous measures were made via nonparametric Wilcoxon 2-sample tests. Categorical measures were compared between noncontracted and contracted groups by χ² tests or Fisher exact tests. Analysis for 95% CIs for differences of proportions or median values in cohorts was performed.

Results

In the contracted group (15 patients), the mean age was 62 years, and 12 were male; in the control group (30 patients), the mean age was 58 years, 26 were male. A total of 25 high-risk alcohol misusers were identified and contracted in the study period. Of these, 15 patients had free flap reconstruction and were included in this study. No patients undergoing surgery and presented with a contract refused signing for the intervention. Our control group consisted of 30 serial patients identified undergoing free flaps and reported drinking more than 7 US drinks per day. No statistically significant difference was found between the 2 cohorts for any demographic, morbidity, or surgical variables (Table 1). No contracted patients had any adverse outcomes at home in the preoperative period. Abstinence was measured based on patient and family reporting. The contracted group had an average of 14.8 days (median, 12 days; range, 7-35 days) of reported abstinence prior to surgery. Abstinence was achieved at home for 11 patients, in a detoxification facility for 2 patients, and with preoperative admission for 2 patients. The detoxification facility was an Ohio State University affiliate that specializes in comprehensive drug and alcohol addiction recovery services. Patients were referred by the social worker and attending surgeon, and all major insurances were accepted. When patients were admitted to the hospital for detoxification, they were either admitted under the care of the attending surgeon or admitted under a hospitalist service depending on the patient comorbidities and active issues other than the ethyl alcohol consumption.

Table 1. Preoperative Characteristics and Surgical Variables.

Characteristic	Group, No. (%)
	Contracted (n = 15)	Control (n = 30)
Age mean, y	62	58
Male, %	12 (80)	26 (87)
Preoperative morbidities
Current smoker	15 (100)	30 (100)
Diabetes mellitus	0	1 (3)
Hypertension	10 (67)	13 (43)
Coronary artery disease	7 (47)	6 (20)
COPD	11 (73)	28 (93)
Peripheral vascular disease	5 (33)	7 (23)
Untreated hypothyroidism	1 (7)	3 (10)
Malnutrition	2 (13)	9 (30)
Surgical variables
Recurrent disease	2 (13)	6 (20)
Prior radiation therapy	1 (7)	6 (20)
AJCC stage
I	0	3 (10)
II	2 (13)	1 (3)
III	5 (33)	3 (10)
IV	8 (53)	23 (77)
Free flap reconstruction
RFFF	8 (53)	15 (50)
ALTFF	3 (20)	4 (13)
FFF	4 (27)	9 (30)
Other	0	2 (7)

Abbreviations: AJCC, American Joint Committee on Cancer; ALTFF, anterolateral thigh free flap; COPD, chronic obstructive pulmonary disease; FFF, fibular free flap; RFFF, radial forearm free flap.

In direct comparisons between the groups, the rate of alcohol withdrawal (63% vs 0%; difference, 0.63; 95% CI, 0.33-0.85), delirium (73% vs 0%; difference, 0.73; 95% CI, 0.45-0.92), cellulitis (43% vs 7%; difference, 0.37; 95% CI, 0.04-0.64), and wound dehiscence (67% vs 13%; difference, 0.53; 95% CI, 0.22-0.79) was higher in the noncontracted group vs the contracted group. Furthermore, hospital stay (median 13 days vs 9 days; difference, 5 days; 95% CI, 3-7 days) and time lapse to starting adjuvant radiation therapy (median 60 days vs 42.5 days, difference, 15 days; 95% CI, 5-25 days) was significantly longer in the noncontracted group vs the contracted group (Table 2).

Table 2. Postoperative Complications and Level of Service.

Variable	Group, No. (%)		Difference (95% CI)a
	Contracted (n = 15)	Control (n = 30)
Complications
Myocardial infarction	0	3 (10)	0.10 (−0.22 to 0.41)
Cerebrovascular accident	1 (7)	2 (7)	0 (−0.32 to 0.32)
Cellulitis	1 (7)	13 (43)	0.37 (0.04 to 0.64)
Sepsis	0	6 (20)	0.20 (−0.13 to 0.50)
Alcohol withdrawal	0	19 (63)	0.63 (0.33 to 0.85)
Delirium	0	22 (73)	0.73 (0.45 to 0.92)
Death	0	1 (3)	0.03 (−0.29 to 0.35)
Fistula	1 (7)	10 (33)	0.27 (−0.06 to 0.56)
Wound dehiscence	2 (13)	20 (67)	0.53 (0.22 to 0.79)
Return to operating room (<30 d)	0	4 (13)	0.13 (−0.19 to 0.44)
Length of stay, median (25th to 75th percentile), d	9 (8 to 9)	13 (11 to 15)	5 (3 to 7)b
Time to begin adjuvant radiation therapy, median (25th to 75th percentile), d	43 (40 to 59)	60 (60 to 65)	15 (5 to 25)B

^aDifference (95% CI): Difference = Proportion in Control group − proportion in Contracted Group.

^bDifference (95% CI): Difference = Median in Control Group − Median in Contracted Group.

Further analysis of all wound dehiscences in the control group showed that a predominance was a result of donor site disruption from uncontrolled extremity movement during withdrawal (14 of 20 patients [70%]). While patients were restrained and/or sedated during active withdrawal, we found that damage to donor sites occurred prior to the initiation of restraints or deep sedation.

Discussion

This study marks the first report of a preoperative abstinence protocol for patients with head and neck cancer. Patients represented in this study are high-risk surgical candidates given the abundance of comorbidities, length of surgery, and required length of stay postoperatively (Table 1). Most striking is that all patients identified in the control and contracted groups were active smokers, and more than 85% of each cohort presented with advanced-stage disease. This is important because it shows that our abstinence protocol resulted in a significant reduction in alcohol withdrawal, delirium, surgical site infection, and wound dehiscence compared with the control group despite advanced disease, smoking, and long surgical procedures.

We believe the reduction in surgical site infection in the contracted group was likely due to improved immune function and tissue healing after a period of abstinence, as is supported by the literature. Beyond reducing inpatient complications, length of stay was significantly improved in the contracted group. This difference is secondary to the additional hospitalization required by patients in active withdrawal and associated complications. Interestingly, the effects of the protocol seem to extend beyond the inpatient stay, because time to adjuvant treatment was reduced in the contracted group. This is believed to be a function of improved wound healing and increased patient compliance in keeping postoperative appointments in the contracted group. The contracted group had a significant reduction in time to radiation therapy, which has been shown to reduce locoregional failure in advanced head and neck cancer.

In Denmark, a randomized clinical trial was performed in which alcohol misusers undergoing colorectal surgery were randomized to having a 1-month intervention of abstinence from alcohol vs no intervention. Abstinence was achieved by administering disulfiram, an aldehyde dehydrogenase inhibitor resulting in significantly less postoperative complications (31% vs 74%), myocardial ischemia (23% vs 85%), and arrhythmias (33% vs 86%). While the ethics of giving patients an agent that will force them to be intolerant of alcohol (eg, an aldehyde dehydrogenase inhibitor) are controversial, it is clearly effective and remains a consideration going forward. We chose a different strategy centered on patient and family empowerment while achieving the goal of cessation preoperatively. The protocol is bolstered by significant education at a time in which the patient has important decisions while facing a life-threatening illness. It is critical that clinicians use patient personal responsibility and preventative medicine to attenuate costs and improve outcomes, particularly at a time in which physicians are held accountable for quality metrics. This study shows that empowering a patient with this alcohol abstinence contract can significantly improve quality metrics even without objective testing.

Limitations

There are several limitations of this study, including the retrospective nature of data collection, the selection bias introduced with physician’s discretion and patients who agree to sign the contract, and the lack of objective methods to monitor patient compliance (point-of-care breathalyzer or serum alcohol laboratories). Patient alcohol testing was not performed owing to a lack of resources and funding to provide such devices. In addition, the protocol does require coordinated efforts and resources of many clinicians and ancillary staff. We did not measure the differences in preoperative clinic times and resource utilization as this data are not available. Further study is needed to determine the costs associated with this intervention. Clearly, there are greater upfront costs, especially with those patients requiring preoperative admission. While it seems intuitive that the reduction in postoperative complications and level of service would lead to an overall reduction in costs, we have not done this analysis on this limited cohort and cannot draw any definitive conclusions. We believe that our data are compelling and show a significant impact that a preventative protocol can have even without a standardized objective measurement of usage precontract and postcontract. Despite this, we feel that performing a future prospective study in a larger cohort that also queries patients for preconceived opinions on alcohol cessation and patient responsibility would help reduce selection bias and help define which subsets of patients may need additional interventions for success.

Conclusions

An alcohol abstinence program for surgically managed patients with head and neck cancer requiring free flap reconstruction is safe. These early results are encouraging because they suggest that an alcohol abstinence program could reduce morbidity and improve outcomes.