Alam M, Ibrahim O, Nodzenski M, Strasswimmer JM, Jiang SIB, Cohen JL, Albano BJ, Batra P, Behshad R, Benedetto AV, Chan CS, Chilukuri S, Crocker C, Crystal HW, Dhir A, Faulconer VA, Goldberg LH, Goodman C, Greenbaum SS, Hale EK, Hanke CW, Hruza GJ, Jacobson L, Jones J, Kimyai-Asadi A, Kouba D, Lahti J, Macias K, Miller SJ, Monk E, Nguyen TH, Oganesyan G, Pennie M, Pontius K, Posten W, Reichel JL, Rohrer TE, Rooney JA, Tran HT, Poon E, Bolotin D, Dubina M, Pace N, Kim N, Disphanurat W, Kathawalla U, Kakar R, West DP, Veledar E, Yoo S. Adverse Events Associated With Mohs Micrographic SurgeryMulticenter Prospective Cohort Study of 20 821 Cases at 23 Centers. JAMA Dermatol. 2013;149(12):1378-1385. doi:10.1001/jamadermatol.2013.6255
Detailed information regarding perioperative risk and adverse events associated with Mohs micrographic surgery (MMS) can guide clinical management. Much of the data regarding complications of MMS are anecdotal or report findings from single centers or single events.
To quantify adverse events associated with MMS and detect differences relevant to safety.
Design, Setting, and Participants
Multicenter prospective inception cohort study of 21 private and 2 institutional US ambulatory referral centers for MMS. Participants were a consecutive sample of patients presenting with MMS for 35 weeks at each center, with staggered start times.
Mohs micrographic surgery.
Main Outcomes and Measures
Intraoperative and postoperative minor and serious adverse events.
Among 20 821 MMS procedures, 149 adverse events (0.72%), including 4 serious events (0.02%), and no deaths were reported. Common adverse events reported were infections (61.1%), dehiscence and partial or full necrosis (20.1%), and bleeding and hematoma (15.4%). Most bleeding and wound-healing complications occurred in patients receiving anticoagulation therapy. Use of some antiseptics and antibiotics and sterile gloves during MMS were associated with modest reduction of risk for adverse events.
Conclusions and Relevance
Mohs micrographic surgery is safe, with a very low rate of adverse events, an exceedingly low rate of serious adverse events, and an undetectable mortality rate. Common complications include infections, followed by impaired wound healing and bleeding. Bleeding and wound-healing issues are often associated with preexisting anticoagulation therapy, which is nonetheless managed safely during MMS. We are not certain whether the small effects seen with the use of sterile gloves and antiseptics and antibiotics are clinically significant and whether wide-scale practice changes would be cost-effective given the small risk reductions.
Mohs micrographic surgery (MMS) offers the promise of skin cancer excision with the highest cure rates and greatest degree of tissue sparing. Like many dermatologic procedures performed under local anesthesia, MMS is also reported to be very safe, with a low incidence of intraoperative and postoperative adverse events.1- 15
Much of the data pertaining to the complications of MMS are anecdotal, are from single centers, or concern a single type of adverse event.2- 14 More recently, at least 1 impressive effort to combine safety data from several centers has been undertaken.1 Without exception, these previously published studies confirm that MMS is associated with a low risk for minor complications and a rare risk for serious adverse events.
However, the data pertaining to the safety of MMS remain limited in scope. In particular, single-center studies and the few multicenter studies that have been performed tend to include surgeons who trained together or at the same centers and hence may not be representative of the typical MMS surgeon. In addition, given the very low rates of adverse events associated with MMS, even larger numbers of cases than have been examined at present must be evaluated to provide reliable incidence data for specific complications and to make statistical comparisons. Finally, typical multicenter studies of MMS have not included robust mechanisms to remind record keepers, to obtain and preserve data through a secure computerized database, and to conduct interim analyses to test for data entry errors and encourage uniformity in data collection.
The purpose of this multicenter prospective cohort study was to better assess the safety of MMS by addressing and overcoming some of the shortcomings of the earlier studies. Specifically, many more cases were studied from disparate and unrelated centers, mechanisms were implemented to improve data quality, and specific adverse events were analyzed in greater depth.
We conducted a multicenter prospective cohort study involving MMS centers from different geographic locations in the United States. Data acquisition occurred for 8 months per center, with the start dates staggered across 8 months to minimize bias associated with seasonal variation in patient presentation. Data collection was accrued directly on the DermBase Web-based interface. Study procedures were performed under the authorization of 2 expedited Northwestern University institutional review board (IRB) protocols, which collectively authorized data collection and managed transmittal of de-identified data across institutions to enable analysis. All data collected were de-identified, and informed consent was waived per IRB direction.
Centers were selected to provide geographic diversity and to provide technique diversity because efforts were made to avoid recruiting multiple physicians trained at the same institutions. All physicians were members of the American College of Mohs Surgery. Key nonphysician study personnel identified by the lead physician at each center were trained on data acquisition and reporting by personnel (M.A. and O.I.) at the coordinating center (Northwestern University). Data pertaining to numbers and types of tumors treated, extended data acquisition regarding rare tumors, and detailed information about adverse events were uploaded each working day to the DermBase site from each participating center.
DermBase is a Web-based collaborative research tool initially funded by a grant from the Sulzberger Institute for Dermatologic Education of the American Academy of Dermatology to the current project coordinator (J.M.S.) and another researcher (S.I.B.J.). The mission of DermBase is to facilitate multisite research in clinical dermatology and dermatologic surgery. DermBase provides methods for oversight and critique of study design before study initiation, a simplified pathway for IRB approval for new studies, and tools to simplify data collection and entry.
The multicenter MMS study was initially submitted to the DermBase advisory board in January 2010. The proposed study was then posted on the DermBase site for public comment, comments were collated, and the protocol was approved, at which point it was approved by the DermBase advisory board for initiation by the principal investigator (M.A.). Northwestern University research staff (M.A. and O.I.) worked with DermBase programmers to develop appropriate data collection interfaces and test these before study commencement.
Monthly interim analyses were performed during data collection to cross-check the integrity of the stored data. At the conclusion of data collection for this study, the collated data were downloaded, checked for errors, and delivered to Northwestern University for analysis.
All data were input directly onto the Web-based interface by participating centers and then downloaded by the coordinating center for analysis. The following 2 principal data evaluation forms were used: (1) the General Practice Questionnaire, which elicited information regarding routine patient management and surgical technique practices at each center, and (2) the Weekly Questionnaire (closed weekly but with daily data entry in day-specific columns), which recorded the numbers of tumors treated by type and date. In addition, an adverse event reporting form including detailed clinical information was completed when any unexpected adverse events were reported.
Covariates collected for all cases were the practice type, practice setting (ie, urban, suburban, or rural), practice location (ie, geographic location), median years of experience by the physician, use of sterile gloves during MMS/reconstruction, type of sterile preparation used, use of separate surgical trays for MMS/reconstruction, use of antibiotic prophylaxis, use of cautery, and postoperative wound care regimen. Additional covariates collected for patients with adverse events included patient sex, patient age, anatomic location of the tumor, histologic type of the tumor, presence of implanted appliances, immunosuppression, anticoagulant use, tobacco use, history of diabetes mellitus, type of surgical repair, type of postoperative bandage, use of a hemostatic agent, postoperative tumor length and width, postoperative tumor depth, presence of signs or symptoms of infection, culture positivity for infection, organism type if cultured, presence of persistent postoperative bleeding, hematoma, deep vein thrombosis, dehiscence, persistent dysesthesia, motor nerve injury, hypertrophic scar/keloid, necrosis, ectropion, eclabium, occurrence of postoperative hospitalization, myocardial infarction, stroke, pulmonary embolism, death, and other serious adverse events. Medical problems, medications used, age, sex, personal and family history of skin cancer, social history, and history of radiotherapy were not routinely recorded for patients unless adverse events were noted.
This study was not designed to assess long-term functional deficits associated with MMS. Infection was assessed by clinical impression and culture positivity. During data analysis, either method was deemed sufficient to label the case as infected.
All study sites were prospectively asked to track all complications actively for the duration of the study. To ensure appropriate ascertainment of adverse events, surgeons were asked to contact patients by telephone within 24 hours of the procedure to query patients regarding any signs, symptoms, and events requiring medical care. In addition, all patients in whom postoperative visits were feasible were followed up at least once within 1 month of surgery (ie, excluding debilitated patients with granulating wounds who lived in nursing homes, patients who were relocating or embarking on prolonged travel, patients who insisted on having their sutures removed or receiving other wound care closer to home by other medical personnel, and patients who declined to return for follow-up or failed to attend their scheduled follow-up visits). At the scheduled follow-up visits, appropriate history was obtained and physical examinations were conducted to elicit the signs and symptoms of any adverse events and any patient reports of transient adverse events that had resolved before presentation. Follow-up visits were generally staffed by study physicians, with nonphysician health care personnel seeing the patients in some cases.
We performed descriptive analyses to show the characteristics of study institutions, including practice type, urban or suburban setting, and regional location. General practices of study institutions were also quantified according to whether a given practice was sometimes, always, or never performed. Similar analysis was performed to characterize the breakdown of tumors treated in this study by tumor type and to provide demographic information of patients who experienced adverse event complications. Presented patient demographics included age, sex, the most common tumor type, and the most common anatomic location of the tumor.
All adverse events were characterized according to their associations with known risk factors and intraoperative procedures. The frequencies of each of the 9 types of individual adverse events were found and collapsed into the following 4 subtypes of adverse events: bleeding, infection, impaired wound healing, and serious adverse events. Patient demographic information was reported for each subtype, and the bleeding, infection, and impaired wound healing subtypes were further described according to associations with known risk factors.
We performed χ2 testing with Yates continuity corrections to assess associations of infectious and impaired wound-healing complications with general surgical practices thought to be linked to these complications. Likewise, χ2 tests with Yates continuity corrections were performed to examine associations between bleeding complications and risk factors for bleeding. Analysis of variance with adjustment for multiple pairwise testing was used to ascertain any difference in age. To explore associations detected by χ2 tests, we used Poisson regression to create multiple models to further assess the significance of expected predictors of adverse events.
For the 70-week period from November 29, 2010, through April 2, 2012, 23 centers for MMS (including 36 physicians) in the United States prospectively collected data for a mean of 35 consecutive weeks. A total of 20 821 tumors were studied (Table 1). One hundred forty-nine adverse events were reported, for an adverse event rate of 0.72%. Participating MMS surgeons were removed from fellowship training by a median of 12.6 years (range, 4-32 years).
Overall, postoperative visit frequency was consistent across centers, but centers differed substantially in their use of nonphysician health care personnel and midlevel providers as first assessors during such visits. The mean dropout rate for in-person follow-up visits was reported as 4.1%. The frequency of adverse events that occurred more than 1 month after a procedure varied across centers because no protocol-based recommendation for these measures existed. However, because most of the acute adverse events associated with MMS likely manifest within 1 month, we did not consider the lack to be a major cause of underreporting.
Demographic information associated with adverse events, including patient age and sex and the anatomic location and histologic type of the tumor being treated, is reported in Table 2. Adverse events listed by type and frequency, demographic factors associated with the major subtypes of adverse events, and the most common anatomic areas for the major subtypes of adverse events are also recorded in Table 2. Table 3 gives the association of major adverse event subtypes with specific risk factors for these adverse events. Table 4 reports the adverse event rates associated with general surgical practices. Based on Poisson regression analysis, the use of sterile gloves, chlorhexidine gluconate preparation, oral antibiotic prophylaxis, and cautery continued to be independently significantly associated with the risk for infection; the only factor associated with impaired wound healing was the use of cautery (Table 5). Median years of physician experience was not a significant predictor of adverse events (P = .64 for infection, P = .73 for bleeding complications, and P = .38 for impaired wound healing).
Serious adverse events occurred in 4 cases, constituting 2.7% of all adverse events and 0.02% of all procedures. The mean age of patients who experienced serious adverse events was 72.5 years, and 2 (50%) were male. All serious adverse events were hospitalizations, including 1 for signs and symptoms of infection with partial necrosis and 2 for culture-positive infection with methicillin-resistant Staphylococcus aureus. The latter 2 cases were not associated with hemodynamic instability but rather resulted from the failure of oral antibiotic therapy in the context of debilitation and multiple medical problems. The fourth hospitalization was related to the MMS because it was reported as such by the operating MMS surgeon, who noted a site-specific infectious problem managed at an outlying medical center, but additional details are not available. However, available data indicate that none of the hospitalized patients had temporally related stroke, myocardial infarction, pulmonary embolus, other life-threatening events, or death.
This prospective cross-sectional study of 20 821 cases of MMS performed by 36 surgeons at 23 US centers during a 70-week period detected an adverse event rate of 0.72%, which is toward the lower end of the range of 0.5% to 2.0% reported in the literature. The most common adverse events reported were infections at 61.1% of complications, followed by dehiscence and partial or full necrosis (20.1%) and bleeding and hematoma (15.4%). Four adverse events (2.7%) were serious, and all resulted in hospitalization but not death.
Although complications were most commonly associated with procedures on the face, the legs and scalp were common anatomic sites for infectious complications and the neck was a common site for bleeding complications, with wound-healing complications most frequent on the nose and to a lesser extent on the scalp and ears. The mean age of patients with complications exceeded 70 years, and half of these patients were male. Most of the patients with postoperative bleeding were undergoing anticoagulation therapy, with 1 (5%) known to have no anticoagulant use.
This study is, to our knowledge, the first of sufficient size to document a potential association between use of sterile vs clean gloves during MMS and a reduced risk of infection (P = .04).2 Most of the MMS surgeons surveyed used sterile gloves for reconstruction, but fewer used them during MMS. Whether the use of sterile gloves for MMS may reduce the risk for infection is difficult to ascertain from this observational study owing to possible confounders inherent in such a study design. A randomized clinical trial would be needed to validate any causal association.
Preoperative use of the antiseptic chlorhexidine (P < .001) and perioperative use of oral and/or topical antibiotics (P < .001) were also associated with a lower risk for infection. Although preparation with chlorhexidine is typical, routine antibiotic prophylaxis is more controversial given the risks to the patient and the risks for antibiotic resistance. Even if the detected associations were suggestive of causality, which cannot be shown by this study, a single method of skin decontamination may be sufficient, and implementation of several methods may not increase the aggregate benefit.
Excessive use of cautery is believed to devitalize tissue, thus inducing suboptimal wound healing and even infection, and we found that cautery was associated with impaired wound healing (P < .001). However, bleeding complications would have been more likely in selected patients at high risk for bleeding if cautery were withheld.
Overall, the strengths of this study include the very large number of MMS cases, the prospective design, and the multicenter approach, which has been undertaken by only one other group. Multiple surgeons trained by the same fellowship directors were excluded to ensure heterogeneity in practice patterns. Online Web-based data entry and archiving ensured data integrity, and the quality of the data entry was checked weekly by a data-entry expert (N.P.) and a coinvestigator (I.O.).
Findings of this study confirmed earlier data that serious adverse events associated with MMS are exceedingly rare. Although Kimyai-Asadi et al5 reported 1 case of such an adverse event in 3937 consecutive cases of MMS in a single center, we found no cases of myocardial infarction, stroke, pulmonary embolus, or death in a larger sample.
Compared with prior studies, our study indicated a lower total adverse event rate of 0.72% per case. Previously, one single-center study of 1343 cases reported a rate of 1.64%6 encompassing mostly bleeding episodes. One multicenter study found a 2.6% rate of minor complications in 1792 cases.1
Surgical-site infections were also less frequent in our study than in past reports. Historic rates have included 0.91% in 1204 cases4 and 2.29% in 1047 cases (530 MMS and 517 excisions).9 Infections were reportedly reduced from 2.5% to 0.2% in 908 cases by intraincisional injected antibiotics.7
In this study, flaps and grafts were risk factors in one-half to two-thirds of bleeding, infection, and wound-healing complications. Because the underlying rate of flaps and grafts after MMS is much lower than half—probably closer to 37%16—we can conclude that flap and graft repairs are disproportionately likely to be associated with complications. We are unclear whether this risk is a function of the repair type or of some covariate associated with flap or graft use, such as anatomic location or defect size. For instance, more than two-thirds of bleeding complications occurred when final wound defects were longer than 1.5 cm.
The curious observation that bleeding is associated with reduced infection may be distantly related to the finding that topical antibiotic application after surgery reduces the risk for infection. That is, bleeding and active cleansing may mobilize potentially infective debris away from the healing wound. Patients developing infections were more evenly split between men and women than were patients with bleeding and wound-healing problems, who were more likely to be male (P = .02).
Risk factors associated with all adverse events included current and past smoking. Surprisingly, twice as many adverse events were associated with a history of smoking vs current smoking. Possibly, current smokers are reluctant to confess to this habit and instead characterize themselves as former smokers. Surgeons may need to expect wound-healing issues in “former” smokers.
This study has limitations. Cases and adverse events could have been underreported. However, to minimize this, methods to optimize ascertainment during the study period included training and retraining of data entry staff at each center, electronic capture of cases, checking of data entry by third parties, and periodic reminders to staff and physicians. Another limitation is that the study does not include all MMS cases performed in the United States. Despite being the largest, most heterogeneous sample of sites and cases to date, this study is still at risk for selection bias. Finally, the incidence of infection was classified on the basis of signs and symptoms of infection or on culture isolation of pathogenic organisms. Conceivably, severe inflammation occasionally may have been misclassified as infection; if so, however, the true cumulative infection rate would be even lower than the rate detected.
Significantly, the infection-control measures suggested by this study are associated with only very small clinical benefits in infection reduction. Thus, although preparation with chlorhexidine, antibiotic prophylaxis, or the use of sterile gloves during MMS may reduce postoperative infections, the absolute risk reduction for each ranges from 0.45% to 0.53%, and risk reductions are likely not additive. In large studies such as this one, it is possible to detect statistical significance that is associated with only modest clinical significance. In addition, to the extent that our study is observational and not a randomized clinical trial, it is susceptible to the effects of unknown confounding variables. If the infection risks detected in this study were validated in a randomized clinical trial, their absolute values would likely be seen to decrease further.
In this context, whether infection risk reductions of 0.5% or less are considered worth pursuing would depend on a cost-benefit analysis that examined the aggregate cost per infection prevented. For instance, the expense of routine oral antibiotic therapy would be incurred a mean of 200 times to prevent a single infection. Harms can also be considered. That is, some measures used to mitigate infection risk may have other potentially deleterious effects. Routine oral antibiotic prophylaxis, apart from its expense, may predispose to allergy in the individual and antibiotic-resistant organisms in the population. Because MMS is already extremely, remarkably safe, further studies, including randomized clinical trials, are needed to verify the findings of this study before they can be considered the foundation of adjustments, however modest, in clinical practice.
This study provides, to our knowledge, the largest prospectively collected multicenter cohort sample of cutaneous neoplasms undergoing MMS. Significant findings include a very low rate (0.72%) of minor postoperative complications after MMS, an exceedingly low rate of serious adverse events (0.02%) that consisted exclusively of short-term hospitalizations, and a zero rate of death or permanent disability. Although rare, adverse events after MMS are most commonly infections and somewhat less often bleeding or impaired wound healing. Use of sterile gloves, local antiseptic (chlorhexidine), and antibiotics (oral and topical) may be associated with decreased risk of infection, but potential risk reductions are minute and require further corroboration. Most bleeding complications occur in patients receiving anticoagulants or antithrombotics, but most patients receiving blood thinners do not develop complications. Infectious complications are more common in younger and female patients. Flaps and grafts are more likely than other repairs to be associated with complications. Overall, MMS is exceedingly safe, with even lower overall complication rates than the low rates previously reported by smaller and single-center studies. Although future studies may confirm and extend these findings, this study is notable for being adequately powered to provide detailed information about adverse events associated with MMS.
Accepted for Publication: June 18, 2013.
Corresponding Author: Murad Alam, MD, MSCI, Department of Dermatology, Feinberg School of Medicine, Northwestern University, 676 N St Clair St, Ste 1600, Chicago, IL 60611 (firstname.lastname@example.org).
Published Online: September 30, 2013. doi:10.1001/jamadermatol.2013.6255.
Author Contributions: Drs Alam and Veledar and Mr Nodzenski had full access to the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Alam, Strasswimmer, Cohen, Goldberg, Kouba, Dubina, Pace, Kim, Kathawalla, West.
Acquisition of data: Alam, Ibrahim, Strasswimmer, Jiang, Cohen, Albano, Batra, Behshad, Benedetto, Chan, Chilukuri, Crocker, Crystal, Dhir, Faulconer, Goldberg, Goodman, Greenbaum, Hale, Hanke, Hruza, Jacobson, Jones, Kimyai-Asadi, Kouba, Lahti, Macias, Miller, Monk, Nguyen, Oganesyan, Pennie, Pontius, Posten, Reichel, Rohrer, Rooney, Tran, Bolotin, Pace, Kim, Disphanurat, Kathawalla, Kakar, Yoo.
Analysis and interpretation of the data: Alam, Nodzenski, Strasswimmer, Cohen, Goldberg, Poon, Veledar, Yoo.
Drafting of the manuscript: Alam, Ibrahim, Strasswimmer, Jiang, Cohen, Dhir, Faulconer, Kouba, Oganesyan, Reichel, Disphanurat.
Critical revision of the manuscript for important intellectual content: Alam, Nodzenski, Strasswimmer, Cohen, Albano, Batra, Behshad, Benedetto, Chan, Chilukuri, Crocker, Crystal, Goldberg, Goodman, Greenbaum, Hale, Hanke, Hruza, Jacobson, Jones, Kimyai-Asadi, Lahti, Macias, Miller, Monk, Nguyen, Pennie, Pontius, Posten, Rohrer, Rooney, Tran, Poon, Bolotin, Dubina, Pace, Kim, Kathawalla, Kakar, West, Veledar, Yoo.
Statistical expertise: Veledar.
Obtained funding: Alam.
Administrative, technical, and material support: Nodzenski, Strasswimmer, Jiang, Albano, Batra, Chilukuri, Dhir, Faulconer, Goldberg, Goodman, Greenbaum, Hale, Hanke, Jones, Kimyai-Asadi, Miller, Monk, Nguyen, Oganesyan, Pennie, Posten, Poon, Bolotin, Dubina, Pace, Kim, Disphanurat, Kathawalla, Kakar, West.
Study supervision: Alam, Strasswimmer, Yoo.
Conflict of Interest Disclosures: Dr Alam is employed at Northwestern University and has been a consultant for Amway and Leo Pharma, unrelated to this research. Northwestern University has a clinical trials unit that receives grants from many corporate and governmental entities to perform clinical research. Dr Alam has been principal investigator in studies funded in part by Allergan, Medicis, Bioform, and Ulthera. In all cases, grants and gifts in kind have been provided to Northwestern University and not Dr Alam directly, and Dr Alam has not received any salary support from these grants. Dr Alam receives royalties from Elsevier for technical books he has edited (<$5000 per year).
Funding/Support: This study was supported by Departmental Research Funds, Department of Dermatology, Northwestern University. DermBase, the web-based data entry and archiving repository used for this study, was funded by grants from the Sulzberger Institute for Dermatologic Education of the American Academy of Dermatology.
Role of the Sponsor: The Department of Dermatology, Northwestern University, was fully responsible, and Sulzberger Institute for Dermatologic Education had no role: in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.