Antibiotic Use in Lateral Skull Base Surgery: A Survey of the North American Skull Base Society

Abstract

Objectives  Antibiotic use in lateral skull base surgery (LSBS) has not been thoroughly investigated in the literature. There is wide variability in antibiotic use and insufficient data to guide management. This study aims to describe the factors and patterns influencing antibiotic use in LSBS among the membership of the North American Skull Base Society (NASBS).

Design  An online-based survey was designed and distributed to the membership of the NASBS. Data was analyzed using bivariate analysis and logistic regression modeling.

Setting  Online-based questionnaire.

Participants  NASBS membership.

Main Outcome Measures  Use of intraoperative antibiotics and use of postoperative antibiotics.

Results  The survey response rate was 26% (208 respondents). Of the 208 total respondents, 143 (69%) respondents performed LSBS. Most respondents are neurosurgeons (69%) with the remaining being otolaryngologists (31%). The majority of respondents (79%) are fellowship-trained in skull base surgery. Academic or government physicians make up 69% of respondents and 31% are in private practice with or without academic affiliations. Bivariate analysis showed that practice setting significantly influenced intraoperative antibiotic use ( p  = 0.01). Geographic location significantly affected postoperative antibiotic use ( p  = 0.01). Postoperative antibiotic duration was significantly affected by presence of chronic otitis media, cerebrospinal fluid leak, and surgeon training ( p  = 0.02, p  = 0.01, and p  = 0.006, respectively). Logistic regression modeling showed that the motivation to reduce infection significantly impacted postoperative antibiotic use ( p  = 0.03).

Conclusion  This study demonstrates significant variations in intraoperative and postoperative antibiotic use in LSBS among the NASBS membership. Appropriate guidelines for optimal perioperative antibiotic use patterns should be determined with randomized studies in the future.

Introduction

Lateral skull base surgery (LSBS) encompasses a wide range of otoneurosurgical procedures which are typically classified as clean procedures. ¹ Despite being clean procedures, the severity of potential postoperative intracranial infections following LSBS, such as meningitis, has led many otolaryngologists and neurosurgeons to routinely administer intraoperative and postoperative antibiotics. ¹ Reported rates of postoperative meningitis after LSBS vary greatly in the literature, ranging from 0.14 to 9.9%. ^{1 2 3 4 5}

Although there is a general consensus promoting the use of perioperative antibiotics, there is a paucity of literature on the ideal initiation time and duration of antibiotic administration for LSBS. Patel et al, in their systematic review, examined two otoneurosurgical case series and three meta-analyses on lateral skull base craniotomies and ultimately recommended administration of a single intraoperative antibiotic dose for “clean” LSBS. This recommendation is similar to that from a 2013 study by Bratzler et al. ^{1 6} In the current infectious disease landscape, routine antibiotic administration must be weighed against ethical and medical consequences such as potentially contributing to the development of antibiotic-resistant bacterial strains and direct patient morbidity such as Clostridium difficile enterocolitis. ⁷ To date, most evidence on appropriate perioperative antibiotic use has not come from well-controlled or randomized studies and, as a result, the establishment of evidence-based guidelines has been compromised resulting in widely varying practices around the world. ⁸

In light of the increasing demand for antibiotic stewardship, growing health care costs, and potential side effects of antibiotics, it is necessary to develop proper evidence-based antibiotic prescribing patterns. ⁹ To investigate the extent of variability in antibiotic prescribing in LSBS and the reasons for these differences, we surveyed the membership of the North American Skull Base Society (NASBS).

Methods

A 22-item online-based SurveyMonkey (San Mateo, California, United States) questionnaire was sent electronically to the general membership of the NASBS (Appendix 1). The survey remained open for a total of 12 weeks with three notification emails sent in the first 8-week period. Participation was voluntary and responses were anonymous with no identifying information recorded.

Survey responses were stratified based on whether surgeons performed anterior skull base surgeries (ASBS), LSBS, or both. Respondent demographics such as type of clinical practice, geographic location, years in practice, neurosurgery versus otolaryngology residency, and completion of a skull base surgery fellowship were recorded. Practice volume was assessed using the number of LSBS cases performed per year. Both intraoperative and postoperative antibiotic use was determined, and responses were stratified based on frequency. The survey also included opportunities to select one or more reasons from a list as to why perioperative antibiotics were prescribed, and whether patient comorbidities or surgical factors (i.e., use of lumbar drains) affected antibiotic use ( Table 3 ). Lastly, the type of antibiotic, the duration of use, and method of postoperative infection diagnosis were assessed.

Table 3. Factors influencing intraoperative and postoperative antibiotic prescriptions.

	Intraoperative antibiotics		p -Value	Postoperative antibiotics		p -Value	OR (95% CI)
	n	%		n	%
Patient or surgical factors a
General active infection/Purulence	27	20	0.14	20	22	0.60
Chronic otitis media	13	10	0.56	6	7	0.37
Active CSF leak	19	14	0.43	15	17	0.13
Reduce risk of postoperative infection	118	89	0.18	74	83	0.03	3.8 (1.1–12.9)
Reduce risk of bacteremia	29	22	0.47	15	17	0.13
Use of absorbable packing	9	7	0.96	6	7	0.07
Use of autologous tissue (e.g., abdominal fat, fascia lata)	17	13	0.32	9	10	0.53
Use of nonautologous tissue (e.g., alloderm, DuraGen)	15	11	0.60	12	13	0.19
Use of nonorganic matter (e.g., hydroxyapatite cement, metal hardware)	15	11	0.60	7	8	0.74
Use of lumbar drain	18	14	0.94	15	17	0.92
How I was taught	22	17	0.22	14	16	0.62
Comorbidities that may change prescribing decision b
No, I do not change my prescribing pattern based on comorbidities	112	85	0.0006	67	75	0.09	8.1 (2.4–26.7)
HIV/AIDS	10	8%	0.0005	11	12	0.13	0.1 (0.02–0.3)
Cystic fibrosis	3	3	0.23	4	4	0.09
Diabetes mellitus	14	11	0.0003	20	22	0.17	0.1 (0.03–0.3)
Transplant patient	13	10	0.02	16	18	0.44	0.2 (0.05–0.8)
Cardiac disease	6	5	0.09	6	7	0.07
Pulmonary disease	8	5	0.02	6	7	0.07	0.2 (0.03–0.8)

Abbreviations: AIDS, acquired immunodeficiency syndrome; CI, confidence interval; CSF, cerebrospinal fluid; HIV, human immunodeficiency virus; OR, odds ratio.

Note: Survey responses of patient and surgical factors that influence intraoperative and postoperative antibiotic prescriptions. Bolded p -values and corresponding OR (95% CI) are significant and are results from a logistic regression, which models frequency among prescribers of antibiotics.

^aThe model for patient or surgical factors predicts the probability that the surgeon prescribes antibiotics more frequently than the majority of other surgeons (prescribe antibiotics > 30% of the time).

^bThe model for comorbidities predicts the probability that surgeon prescribes antibiotics 100% of the time.

Statistical analysis was only performed on complete surveys using SPSS Statistics v21.0 (IBM, Armonk, New York, United States). As in our previous study, ¹⁰ bivariate analysis—chi-squared or Fisher's exact test—was used to detect associations between variables. Pairwise comparisons were used to further assess significant values. We used logistic regression modeling to examine the association between patient and surgical factors with both intraoperative and postoperative antibiotic use patterns. The outcome variable was dichotomized based on surgeon prescribing habits, namely whether postoperative antibiotics were prescribed “100% of the time” or “not 100% of the time.” All patient and surgical factors with a p -value of < 0.20 were evaluated for model inclusion using forward selection, and those significantly associated with the outcome were reported in the final model. Statistical significance was set at the conventional α of 0.05.

Results

Respondent Characteristics

Out of 813 NASBS members, 208 (26%) responded. Of these 208 respondents, a total of 143 (69%) members performed LSBS and were included in this study. Most respondents are neurosurgeons ( n  = 98, 69%) with the remaining being otolaryngologists ( n  = 45, 31%). The majority of respondents ( n  = 113, 79%) are fellowship-trained in skull base surgery. Practice setting was assessed and academic or government physicians comprised 69% ( n  = 99) of respondents, and private practice with or without academic affiliation comprised 31% ( n  = 44) of respondents. Responses were received from every continent except Antarctica, with 61% ( n  = 84) of respondents from the United States or Canada, 15% ( n  = 21) from Europe, 16% ( n  = 22) from Asia or Australia, 7% ( n  = 10) from Central or South America, and 1% ( n  = 1) from Africa.

Practice volume was determined using number of cases performed per year and percentage of practice. Of the survey respondents, the largest percentage of surgeons performed between 11 and 25 or 26 and 50 cases annually ( n  = 39, 27% each) and had LSBS comprise 25 to 50% of their practice ( n  = 50, 35%). The majority of respondents performed both endoscopic and open skull base procedures ( n  = 99, 69%), with fewer performing just open procedures ( n  = 38, 27%), and a small minority performing only endoscopic procedures ( n  = 6, 4%) ( Table 1 ).

Table 1. Surgeon characteristics.

	Overall number	Overall percentage	Intraoperative antibiotics		p -Value	Postoperative antibiotics		p -Value
			n	%		n	%
Total respondents	143	100	132	92		89	62
Primary specialty					0.10			0.99
Otolaryngology	45	31	44	98		28	62
Neurosurgery	98	69	88	90		61	62
Fellowship training					0.81			0.57
Yes	113	79	104	92		69	61
No	30	21	28	93		20	67
Years in practice					0.84			0.17
0–5	35	24	31	89		27	77
6–10	35	24	33	94		23	66
11–15	20	14	18	90		12	60
16–20	13	9	12	92		7	54
20+	40	28	38	95		20	50
Number of skull base cases					0.14			0.79
0–10	14	10	14	100		10	71
11–25	39	27	34	87		22	56
26–50	39	27	34	87		26	66
51–100	31	22	31	100		18	58
100+	20	14	19	95		13	65
Skull base surgery as a percentage of practice					0.71			0.85
< 25%	44	31	41	93		26	59
25–50%	50	35	46	92		32	64
51–75%	38	27	34	89		23	61
76–100%	11	8	11	100		8	72
Type of skull base procedures					0.77			0.44
Endoscopic	6	4	6	100		5	83
Open	38	27	35	92		25	66
Both	99	69	91	92		59	60
Work setting					0.01			0.08
Academic or government	99	69	95	96		57	58
Private with or without academic affiliation	44	31	37	84		32	32
Geographic location	137 a	96			0.57			0.01
United States and Canada	83	61	78	93		52	62
Europe	21	15	20	95		7	33
Asia and Australia	22	16	21	95		18	82
Africa	1	1	1	100		1	100
South and Central America	10	7	8	80		8	80

Note: Survey responses for surgeon training and practice characteristics. Percentages may not add up to 100% due to rounding. Bolded p -values are statistically significant.

^aOnly 137 out of 143 respondents reported geographic location.

Intraoperative Antibiotic Use

Among the survey respondents, almost all ( n  = 132, 92%) prescribe intraoperative antibiotics for LSBS. A majority of respondents indicated that they prescribe intraoperative antibiotics 100% of the time ( n  = 118, 83%), with a minority prescribing them “less than 30% of the time,” “30 to 69% of the time,” or “70 to 99% of the time” ( Table 2 ). The most commonly prescribed antibiotics are nonextended spectrum first- and second-generation cephalosporins (e.g., cefazolin, cefuroxime) with 70% ( n  = 93) of surgeons prescribing them. Less commonly used intraoperative antibiotics include third through fifth generation cephalosporins (e.g., cefotaxime, cefepime) and glycopeptides (e.g., vancomycin) ( Table 2 ). When asked about the reason for intraoperative antibiotic use, surgeons most commonly cited “to reduce the risk of postoperative infection” ( n  = 118, 89%). Most physicians responded that they would not change their antibiotic prescribing patterns in the presence of patient comorbidities, such as human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS), cystic fibrosis, diabetes mellitus (DM), etc. ( n  = 102, 83.6%) ( Table 3 ).

Table 2. Antibiotic use patterns.

	Intraoperative antibiotics		Postoperative antibiotics
	n	%	n	%
Intraoperative antibiotic use	132	92
Postoperative antibiotic use			89	62
Frequency of antibiotic use
Never (0%)	11	8	54	38
Infrequent (1–29%)	3	2	4	3
Sometimes (30–69%)	6	4	4	3
Often (70–99%)	5	4	9	6
Always (100%)	118	83	72	50
Antibiotic class prescribed
Nonextended spectrum first to second generation cephalosporin	93	70	64	72
Extended spectrum third through fifth generation cephalosporin	30	23	24	27
Penicillins	11	8	7	8
Anti-pseudomonal penicillins	3	2	2	2
Aminoglycosides	9	7	5	6
Quinolones	1	1	4	5
Macrolides	0	0	1	1
Lincosamides	11	8	7	8
Nitroimidazole	3	2	4	4
Folate inhibitors	2	2	2	2
Glycopeptides	30	23	24	27
Tetracyclines	1	1	0	0
Carbapenems	3	2	2	2

Note: Survey responses for class and frequency of antibiotics prescriptions. Column percentages exceed 100% as multiple answers could be selected.

Our analysis evaluated associations between work setting and use of intraoperative antibiotics. In a multivariable logistic regression model, controlling for surgeon volume, years of experience, geographic location, and type of skull base surgery (open, endo, both), there was a nonsignificant trend toward an association between private practice setting and use of more intraoperative antibiotics (odds ratio [OR] 4.0 [95% confidence interval [CI] 0.9–18.4]; p  = 0.08). After stratifying prescribing frequency into three categories (never, sometimes [< 100%], and always [100%]) and controlling for the aforementioned variables in a cumulative logit regression model, there was a significant association between private practice setting and use of intraoperative antibiotics. Indeed, private practice surgeons prescribed almost four times the odds more intraoperative antibiotics than surgeons in other practice settings (OR 3.7 [95% CI 1.3–10.9]; p  = 0.01) ( Table 5 ). In contrast, we found no relationship between intraoperative antibiotic use and geographic location, surgeon experience, or surgical caseload ( Table 1 ). Our regression model showed that infection status, reconstruction materials, and use of a lumbar drain, among other factors, did not significantly influence intraoperative antibiotic administration ( Table 3 ).

Table 5. Cumulative logit model for frequency of antibiotic use.

	Frequency	Percent	Cumulative frequency	Cumulative percent
Intraoperative antibiotic frequency
Never	11	7.69	11	7.69
Sometimes (< 100%)	14	9.79	25	17.48
Always (100%)	118	82.52	143	100.00
Postoperative antibiotic frequency
Never	54	37.76	54	37.76
Sometimes (< 100%)	27	11.89	71	49.95
Always (100%)	72	50.35	143	100.00

Note: Cumulative logit model for frequency of antibiotic prescription patterns – p (prescribing more antibiotics).

We did find, however, that patient comorbidities changed intraoperative antibiotic prescription patterns. Among surgeons who prescribed intraoperative antibiotics, surgeons who reported “No, I do not change my prescribing pattern based on comorbidities” were more likely to prescribe intraoperative antibiotics all the time (OR 8.1 [95% CI 2.4–26.7]; p  = 0.0006). Similarly, among surgeons who prescribe intraoperative antibiotics, surgeons who reported that comorbidities do influence their decision were less likely to prescribe intraoperative antibiotics all the time. HIV/AIDs, DM, transplant status, and pulmonary disease all reduced the likelihood of intraoperative prescriptions among these surgeons with similar effect size (OR = 0.1). The most common comorbidity to influence surgeon decision was DM, closely followed by transplant and HIV/AIDs.

Postoperative Antibiotic Use

Postoperative antibiotics were used less often than intraoperative antibiotics ( n  = 89, 62%). The most commonly used postoperative antibiotics were first and second generation nonextended spectrum cephalosporins ( n  = 64, 72%), followed by third through fifth generation cephalosporins and glycopeptides ( n  = 24, 27% each). The frequency of postoperative antibiotic administration varied among respondents, with the majority of respondents prescribing postoperative antibiotics “100% of the time” ( n  = 118, 83%). A minority of surgeons use them “70 to 99% of the time” and very few prescribe postoperative antibiotics less than “70 to 99% of the time” ( Table 2 ). Most respondents prescribed postoperative antibiotics to prevent infection ( n  = 74, 83%) or for active infections ( n  = 20, 22%) ( Table 3 ). Surgeons were asked about their symptom criteria for the diagnosis of postoperative infection following LSBS. The most common criteria for diagnosing postoperative infections were “systemic symptoms” ( n  = 81, 91%) and “wound-related issues” ( n  = 80, 90%). “New or worsening headache,” “altered mental status,” and “ear drainage” also appeared to be used frequently for diagnosis of a postoperative infection ( Table 4 ).Bivariate analysis demonstrated geographic region played a significant role in the use of postoperative antibiotics ( p  = 0.01) ( Table 1 ). Our multivariable logistic regression model, controlling for setting (private practice vs. academic/government), surgeon volume, years of experience, and type of skull base surgery (open, endo, both), showed that the association holds that European surgeons are prescribing fewer postoperative antibiotics (OR 0.31 [95% CI 0.12–0.97]; p  = 0.04) compared with the U.S. or Canada and compared with other geographical regions (OR 0.12 [95% CI 0.03–0.51]; p  = 0.004). We used our cumulative logit model and noted similar associations as our multivariable logistic regression model ( Table 5 ). Controlling for other factors, European surgeons prescribe fewer postoperative antibiotics compared with the U.S. or Canada (OR 0.2 [95% CI 0.1–0.7]; p  = 0.01) and compared with other countries (OR 0.1 [95% CI 0.03–0.5], p  = 0.004).

Table 4. Postoperative infection diagnosis criteria.

	n	%
Symptom criteria for postoperative infection diagnosis
Systemic symptoms	81	91
Wound-related issues	80	90
New or worsening headache	36	40
Altered mental status	45	51
Changes in hearing	4	5
New or worsening dizziness	3	3
Ear drainage	42	47

Our regression model also demonstrated that surgeons who reported “risk of postop infection” as one of the reasons they prescribe postoperative antibiotics were more likely to prescribe antibiotics 100% of the time (OR 3.8 [95% CI 1.1–12,9]; p  = 0.03; Table 3 ).

Surgeon experience, surgical volume, and work setting played no role in postoperative antibiotic prescriptions. Additionally, logistic regression modeling demonstrated that patient comorbidities such as HIV/AIDS, cystic fibrosis, or diabetes did not factor significantly into postoperative antibiotic prescriptions. There was, however, a trend among surgeons who prescribe postoperative antibiotics that responded, “No, I do not change my prescribing pattern based on comorbidities” to be more likely to prescribe postoperative antibiotics all of the time (OR = 2.7 [95% CI 0.9–8.2]). Also, surgeons who reported specific comorbidities appeared to be slightly less likely to prescribe postoperative antibiotics all of the time, although this was not significant ( Table 3 ).

Duration of postoperative antibiotic use also varied widely between respondents. Postoperative antibiotics were most commonly used for “24 hours” ( n  = 47, 53%), followed by “24 to 72 hours,” “1 week,” and “2 weeks,” in decreasing frequency. None of the respondents used postoperative antibiotics for longer than 2 weeks. Duration of postoperative antibiotic use was further evaluated by bivariate analysis. The presence of chronic otitis media (COM) was found to significantly impact the duration of antibiotics ( p  = 0.02). Specifically, the presence of COM increased the odds by 15 times for prescribing postoperative antibiotics for 1 week as compared with 24 hours (OR 15.3 [95% CI 1.43–164.56]; p  = 0.02). The presence of cerebrospinal fluid (CSF) leak was also significantly associated with postoperative antibiotic duration ( p  = 0.01). The presence of CSF leak increased the odds of prescribing postoperative antibiotics for 1 week versus 24 hours by over 10 times (OR 10.5 [95% CI 2.04–53.93]; p  = 0.006). Interestingly, how a surgeon was trained also significantly affected postoperative antibiotic duration ( p  = 0.006). There were multiple statistical differences between “How I was taught” and how long postoperative antibiotics were prescribed: 24 hours versus 1 to 2 weeks ( p  = 0.006), 24 to 72 hours versus 1 to 2 weeks ( p  = 0.002), and between 1 week and 1 to 2 weeks ( p  = 0.022) ( Table 6 ).

Table 6. Factors influencing duration of postoperative antibiotic prescriptions.

	Postoperative antibiotics n (%)					p -Value
	24 h	24–72 h	1 wk	1–2 wk	> 2 wk
Reason for antibiotic use
Overall	47 (53)	27 (30)	12 (14)	3 (3)	0 (0)
General active infection/Purulence	10 (21)	4 (15)	5 (42)	1 (33)	0 (0)	0.25
Chronic otitis media	1 (2)	1 (4)	3 (25)	1 (33)	0 (0)	0.02 a
Active CSF leak	3 (6)	6 (22)	5 (42)	1 (33)	0 (0)	0.01 b
Reduce risk of postoperative infection	40 (85)	23 (85)	9 (75)	2 (67)	0 (0)	0.56
Reduce risk of bacteremia	7 (15)	6 (22)	1 (8)	1 (33)	0 (0)	0.52
Use of absorbable packing	1 (2)	4 (15)	1 (8)	0 (0)	0 (0)	0.16
Use of autologous tissue	3 (6)	2 (7)	3 (25)	1 (33)	0 (0)	0.09
Use of nonautologous tissue	4 (9)	4 (15)	3 (25)	1 (33)	0 (0)	0.21
Use of nonorganic matter	2 (4)	2 (7)	2 (17)	1 (33)	0 (0)	0.11
Use of lumbar drain	5 (11)	6 (22)	3 (25)	1 (33)	0 (0)	0.27
How I was taught	7 (15)	2 (7)	2 (17)	3 (100)	0 (0)	0.006 c

Note: Bolded p -values are significant. Further significant pairwise comparisons are denoted by superscripts: a – 24 hours versus 1 week ( p  = 0.02); b – 24 hours versus 1 week (0.006); c – 24 hours versus 1–2 weeks ( p  = 0.006), 24–72 hours versus 1–2 weeks ( p  = 0.002), and 1 week versus 1–2 weeks ( p  = 0.022).

Discussion

Antibiotics are routinely prescribed for LSBS, despite having little evidence to guide prescription patterns, and the practice of using prophylactic antibiotics is well reported in the neurosurgical literature. For example, Korinek et al have published multiple studies regarding the efficacy of antibiotic prophylaxis for craniotomies. In a prospective study of 4,578 patients, those who received prophylactic antibiotics had a lower rate of surgical site infection (5.8%) compared with the initial cohort that only received penicillin-based antibiotic prophylaxis for emergent craniotomies (9.7% surgical site infection rate). ¹¹ In a similarly designed series of 6,243 craniotomies, the authors found that although antibiotic prophylaxis did not prevent meningitis, they were effective in preventing surgical incision infections. ^{12 13} However, in direct contrast, a meta-analysis by Alotaibi et al found that prophylactic antibiotics do, in fact, decrease rates of meningitis postcraniotomy, a result corroborated by another meta-analysis by Barker. ^{14 15} Although this evidence ultimately led Patel et al to recommend one perioperative dose of antibiotics for clean skull base procedures, ¹ it is not surprising that the membership of the NASBS has widely varying antibiotic prescription patterns for LSBS, given the variability of conclusions in the peer-reviewed literature on this topic.

Our survey of the NASBS had a 26% response rate, with 69% performing LSBS. This response rate is typical of Web-based surveys, which routinely have reported rates under 20%. ^{16 17} Most of the respondents are fellowship-trained in skull base surgery and work in an academic institution. The surgeons who responded to our survey were fairly evenly split in their years in practice with one-half under 10 years and over one-fourth in practice for 20 years or greater, reflecting both the diversity of respondents and the high level of expertise and training of our respondents. Most of these surgeons have practices that are comprised of one-fourth to three-fourths skull base procedures and most perform both endoscopic and open LSBS. The respondents encompass a wide geographical area, again, speaking to the diversity of respondents and generalizability of our results.

Almost every respondent uses intraoperative antibiotics, and of those, over 80% prescribe them 100% of the time. This indicates that a majority of respondents prescribe intraoperative antibiotics independent of surgical factors. This reflexive prescription pattern could potentially be explained by the fact that over half of our respondents work in an academic setting. Surgeons who work in academic centers tend to have more complex cases referred to them, which may inherently have higher risk of infection. Moreover, Patnaik et al found in a retrospective study that LSBS has significantly higher rates of complications (i.e., CSF leaks and nerve palsy) compared with ASBS and is likely a result of having comparatively complex surgical approaches and higher risk clinical scenarios such as skull base osteomyelitis abutting the petrous internal carotid artery. The authors do note that for less risky procedures, there is no difference between LSBS and ASBS complication rates. ¹⁸ The more perilous LSBS approaches may partially explain the high rate of intraoperative antibiotic prescriptions in patients with LSB pathology. We also found that academic or government physicians use significantly more intraoperative antibiotics than their private practice counterparts. This could also be partially explained by the 17% of respondents who prescribe antibiotics based on “How I was taught.” However, most surgeons (89%) are prescribing intraoperative antibiotics to “Reduce the risk of postoperative infection.”

While many surgeons will reflexively prescribe intraoperative antibiotics, we found that another group of surgeons, who do not prescribe 100% of the time, significantly factor patient comorbidities such as HIV/AIDS, DM, transplant status, and pulmonary disease into their intraoperative antibiotic use. This is likely due to the increased risk of infection associated with these comorbidities and can be reflected in the ORs being less than 1. Our regression model indicates that these surgeons were much more likely to not prescribe intraoperative antibiotics 100% of the time and may therefore be more discriminating in their choice to prescribe intraoperative antibiotics or use more criteria to make their decision. This is likely due to the variation in prescription patterns around the world.

Our respondents most commonly use first or second generation cephalosporins (70%), in line with our previous study on antibiotic use in ASBS and other studies on LSBS. ^{10 12 13 19} First and second generation cephalosporins, such as cefazolin, have been very popular for antibiotic prophylaxis because of their broad-spectrum efficacy (especially against Staphylococcal species), low toxicity, and most importantly, their effectiveness at preventing surgical site infections. ^{12 13 20 21 22} Altogether, it appears that a large majority of the membership of the NASBS prefers to prescribe intraoperative antibiotics, typically first or second generation cephalosporins, and do not deviate from this prescription pattern based on surgical factors, except for comorbidities such as HIV/AIDs or DM, with academic surgeons prescribing statistically more intraoperative antibiotics than their private practice counterparts.

The use of postoperative antibiotics in LSBS has weaker support in the literature. This is evidenced in our survey where only 62% of respondents prescribe postoperative antibiotics, compared with greater than 90% for intraoperative antibiotics. It appears that surgical experience, as measured by years in practice and case load, did not affect postoperative antibiotic use. The type of skull base procedure did not change prescription patterns either. We found that of all the surveyed surgeon characteristics, only geographic location had a significant impact on whether postoperative antibiotics were prescribed. Our regression demonstrated that European surgeons prescribe fewer postoperative antibiotics compared with U.S. or Canadian surgeons and the other geographical regions. While we did not further survey for the reasons underlying the difference in prescribing patterns, a potential explanation for European surgeons prescribing fewer postoperative antibiotics includes prudent antimicrobial usage, adherence to national or European Union prescribing guidelines that limits treatments with weaker evidence of efficacy or limited cost-effectiveness, and continuing education on antibiotic stewardship for all members of the health care team. ²³

Similar to intraoperative antibiotics, the majority of our respondents prescribed postoperative antibiotics for every case and a first- or second-generation cephalosporin was most commonly used. The most common reason our respondents prescribed postoperative antibiotics was to “reduce the risk of postoperative infection.” Postoperative infections were diagnosed primarily based on symptoms indicative of infection—systemic symptoms, wound-related issues, headaches, altered mental status, and ear drainage. Our analysis found that respondents have an almost four times greater odds of prescribing postoperative antibiotics if they are trying to reduce the risk of postoperative infections. Unsurprisingly, this pattern reflects the literature and common practice guidelines, which recommends antibiotic prophylaxis for clean LSBS and treatment for meningitis. ^{1 14 15 24 25} In contrast to intraoperative antibiotics, comorbidities such as HIV/AIDs, DM, and transplant status do not significantly affect postoperative antibiotic prescriptions in surgeons who do not prescribe them 100% of the time. While most respondents indicated they would not change their prescribing patterns based on these comorbidities, this was not statistically significant. There is, to our knowledge, no study that concludes the necessity of either intraoperative or postoperative antibiotics for comorbidities such as HIV/AIDs, DM, etc. and underlies the wide variation in prescribing patterns.

One of the most contentious areas regarding LSBS postoperative antibiotic prescription patterns is the duration of therapy. As could be expected, due to the limited data available to guide therapy, the membership of the NASBS had widely varying postoperative antibiotics prescription patterns. The majority of respondents prescribed antibiotics for 1 week or less, with most of this group prescribing for only 24 hours. We found that the presence of COM and CSF leak significantly increased the odds of using postoperative antibiotics for 1 week. This can be explained by the increased risk of meningitis and intracranial infections in the setting of CSF leak and COM and justifiable precautionary prescription patterns. ^{24 26 27 28 29} Interestingly, there were significant differences in how surgeons prescribed postoperative antibiotics based on how they were taught during their training. This further reflects the lack of guidelines for prescribing postoperative antibiotics, and although Patel et al recommend a single intraoperative antibiotic dose, without future randomized trials, no definitive prescription duration can be given.In our previous study regarding ASBS antibiotic usage in the NASBS, ¹⁰ we investigated similar patient and surgical factors that influenced the prescription of intraoperative and postoperative antibiotics. In this present study, we evaluated LSBS antibiotic prescription patterns in the membership of the NASBS. Both studies found that only work setting and geographic location significantly affected intraoperative and postoperative antibiotic prescriptions, respectively. Unsurprisingly, the frequency and class of antibiotic prescriptions were similar in both studies, likely due to the reflexive prescription of first or second generation cephalosporins. In addition, both studies found the same comorbidities (i.e., HIV/AIDs, DM, transplant status, and pulmonary disease) affected intraoperative antibiotic prescriptions, but no factor affected postoperative antibiotic prescriptions in LSBS. While our two studies are related, our current study differs in key elements specific to LSBS, such as evaluating for the presence of COM or use of autologous, nonautologous, and nonorganic materials for closure, all of which we found to not significantly affect intraoperative or postoperative antibiotic prescriptions. Moreover, postoperative antibiotic prescription duration was specifically analyzed for LSBS, with the presence of COM, CSF leak, and surgeon training all significantly factoring into prescription patterns.

Limitations of this study include a low response rate and resulting limited ability to identify statistically significant factors that affect antibiotic use. While our response rate of 26% is higher than standard Web-based surveys, ¹⁶ this overall low response rate decreases our ability to generalize our results to the remaining 74% of the NASBS membership that did not respond to the survey and may lead to nonresponse bias. Most of our respondents are from North America and may decrease the generalizability of results to surgeons in other countries, especially Africa, that may prescribe antibiotics at lower rates. Moreover, survey responses were in the form of multiple-choice questions, with no responses permitted outside these selections. If an option did not reflect a surgeon's answers accurately, there was no way to report this, and this could potentially introduce a selection bias. Finally, since this survey was sent out and data was collected retrospectively, there is a possibility of introducing a recall bias.

Conclusion

This study demonstrates significant variations in intraoperative and postoperative antibiotic use in LSBS among the membership of the NASBS. While both intraoperative and postoperative antibiotics are frequently used by our respondents, several surgeon-specific, surgical, and patient factors, including practice setting, geographic location, and postoperative infection risk, influence decision making for antibiotic use. Future randomized studies are needed to determine appropriate perioperative antibiotic prescription patterns and duration.

Appendix 1.

Survey

The purpose of this survey is to determine the use of perioperative antibiotics in skull base surgery among skull base surgeons. This survey will take you no longer than 5 minutes to complete. We appreciate your time and contribution.

1. Do you perform skull base surgery?

   1. Y/N

2. Do you perform lateral skull base surgery?

   1. Y/N

3. What type of skull base surgery do you perform?

   1. Endoscopic

   2. Open

   3. Both

4. In what setting do you work?

   1. Full-time academic position

   2. Private practice with academic affiliation

   3. Independent private practice

   4. State or government run practice

5. In what geographic location?

   1. Northeast United States

   2. Southeast United States

   3. Midwest United States

   4. Western United States

   5. Europe

   6. Asia

   7. Africa

   8. Australia

   9. South America

   10. Central America

   11. Canada

6. What is your primary field?

   1. Otolaryngology

   2. Neurosurgery

7. Are you fellowship trained?

   1. Y/N

8. What are you fellowship-trained in?

   1. Rhinology

   2. Neuro-otology

   3. Endoscropic Skull Base

   4. Open skull Base Surgery

   5. Skull Base microneurosurgery

   6. Skull Base Surgery

   7. Head & Neck

9. How many years have you been in practice since completion of your residency or fellowship training?

   1. 0–5

   2. 6–10

   3. 11–15

   4. 16–20

   5. 20+

10. How many skull base cases (anterior or lateral) do you do per year?

    1. 0–10

    2. 11–25

    3. 25–50

    4. 50–100

    5. 100+

11. Do you prescribe intraoperative antibiotics for lateral skull base surgery?

    1. Y/N

12. What is the frequency at which you prescribe intraoperative antibiotics for lateral skull base surgery?

    1. 100%

    2. 70–99% of the time

    3. 30 to 69% of the time

    4. 1–29% of the time

    5. 0% of the time

13. Which class(es) of intraoperative antibiotics do you prescribe for lateral skull base surgery? (Check all that apply)

    1. 1 ^st or 2 ^nd generation cephalosporin (e.g. cefazolin, cefamandole)

    2. 3 ^rd through 5 ^th generation cephalosporin (e.g. cefotaxime, cefipime, ceftaroline, ceftobiprole)

    3. Penicillin (e.g. oxacillin)

    4. Anti-pseudomonal penicillins (e.g.piperacillin-tazobactam)

    5. Aminoglycosides (e.g. gentamicin)

    6. Quinolones (e.g. ciprofloxacin)

    7. Macrolides (e.g. erythromycin)

    8. Lincosamides (e.g. clindamycin)

    9. Nitroimidazole (e.g. metronidazole)

    10. Folate inhibitors (e.g. trimethoprim-sulfamethoxazole)

    11. Glycopeptides (e.g. vancomycin)

    12. Tetracyclines (e.g. doxycycline)

    13. Carbapenems (e.g. meropenem)

14. Why do you give intraoperative antibiotics for lateral skull base surgery? (Check all that apply)

    1. General active infection/Purulence

    2. Chronic otitis media

    3. Active CSF leak

    4. Reduce risk of postoperative infection

    5. Reduce the risk of bacteremia

    6. Use of absorbable packing

    7. Use of autologous tissue (e.g., abdominal fat, fascia lata)

    8. Use of non-autologous tissue

    9. (e.g., alloderm, DuraGen)

    10. Use of non-organic matter (e.g., hydroxyapatite cement, metal hardware)

    11. Use of lumbar drain

    12. How I was taught

15. Does your decision to give intraoperative antibiotics for lateral skull base surgery differ based on the patient's comorbidities? If so, which comorbidities? (Check all that apply)

    1. No, I do not change my intraoperative prescribing pattern based on comorbidities

    2. HIV/AIDs

    3. Cystic Fibrosis

    4. Diabetes Mellitus

    5. Transplant patient

    6. Cardiac disease

    7. Pulmonary disease

16. Do you prescribe postoperative antibiotics for lateral skull base surgery?

    1. Y/N

17. What is the frequency at which you prescribe postoperative antibiotics for lateral skull base surgery?

    1. 100%

    2. 70–99% of the time

    3. 30 to 69% of the time

    4. 1–29% of the time

    5. 0% of the time

18. What duration of postoperative antibiotic treatment do you prescribe for lateral skull base surgery?

    1. 24–72 hours

    2. 72 hours

    3. 1 week

    4. 2 weeks

    5. >2 weeks

19. Which class(es) of postoperative antibiotics do you prescribe for lateral skull base surgery? (Check all that apply)

    1. 1 ^st or 2 ^nd generation cephalosporin (e.g. cefazolin, cefamandole)

    2. 3 ^rd through 5 ^th generation cephalosporin (e.g. cefotaxime, cefipime, ceftaroline, ceftobiprole)

    3. Penicillin (e.g. oxacillin)

    4. Anti-pseudomonal penicillins (e.g.piperacillin-tazobactam)

    5. Aminoglycosides (e.g. gentamicin)

    6. Quinolones (e.g. ciprofloxacin)

    7. Macrolides (e.g. erythromycin)

    8. Lincosamides (e.g. clindamycin)

    9. Nitroimidazole (e.g. metronidazole)

    10. Folate inhibitors (e.g. trimethoprim-sulfamethoxazole)

    11. Glycopeptides (e.g. vancomycin)

    12. Tetracyclines (e.g. doxycycline)

    13. Carbapenems (e.g. meropenem)

20. Why do you give postoperative antibiotics for lateral skull base surgery? (Check all that apply)

    1. General active infection/Purulence

    2. Chronic otitis media

    3. Active CSF leak

    4. Reduce risk of postoperative infection

    5. Reduce the risk of bacteremia

    6. Use of absorbable packing

    7. Use of autologous tissue (e.g., abdominal fat, fascia lata)

    8. Use of non-autologous tissue

    9. (e.g., alloderm, DuraGen)

    10. Use of non-organic matter (e.g., hydroxyapatite cement, metal hardware)

    11. Use of lumbar drain

    12. How I was taught

21. How do you make the diagnosis of postoperative infection for lateral skull base surgery? (Check all that apply)

    1. Systemic symptoms (e.g. fever, increasing WBC)

    2. Wound-related issues

    3. New or worsening headache

    4. Altered mental status

    5. Changes in hearing

    6. New or worsening dizziness

    7. Ear drainage

22. Does your decision to give postoperative antibiotics for lateral skull base surgery differ based on the patient's comorbidities? If so, which comorbidities? (Check all that apply)

    1. No, I do not change my intraoperative prescribing pattern based on comorbidities

    2. HIV/AIDs

    3. Cystic Fibrosis

    4. Diabetes Mellitus

    5. Transplant patient

    6. Cardiac disease

    7. Pulmonary disease

Erratum: An erratum has been published for this article (DOI: 10.1055/s-0042-1759560).