Welcome back to the journal review. This is our September 2026 rundown of Dermatologic Surgery, and we've got four pieces from the Communications section worth your time — a cautionary case report, a practical retrospective on wound dressings, a big real-world look at skin substitute safety, and a technique note on nasal valve support, complete with an editorial commentary. Let's get into it.

First up is a case report titled "Aggressive Squamous Cell Carcinoma During Treatment With Upadacitinib." This is exactly what it sounds like — a single case report, not a controlled study, so no methods section to dissect here, just a clinical narrative and a discussion worth sitting with.

The background is the growing unease around JAK inhibitors and skin cancer risk. Upadacitinib is a second-generation reversible JAK inhibitor, now approved across a widening list of autoinflammatory indications including ulcerative colitis. The signal problem is this: in rheumatoid arthritis and psoriatic arthritis populations, upadacitinib has been associated with higher nonmelanoma skin cancer rates compared with adalimumab or methotrexate, and there's a dose-dependent relationship reported in RA patients. But — and this is the tension the authors set up — the larger meta-analyses, with a median follow-up of only around four years, haven't found a statistically significant bump in nonmelanoma skin cancer risk with JAK inhibitors as a class. The authors point out, reasonably, that four years may just not be long enough; other immunosuppressive regimens have needed five-plus years of follow-up before their true cancer signal becomes visible. So this case is offered as a data point in that gap — an n of one, but a vivid one.

The patient is a 77-year-old man with a quarter-century history of ulcerative colitis, prior basal cell carcinomas treated with Mohs, a meaningful smoking history, and fair skin — so not a blank slate, he's already got baseline risk factors. He'd been transitioned from mycophenolate mofetil to upadacitinib about five months before this all began. He then develops a well-differentiated SCC on the nasal ala, cleared in a single Mohs stage — unremarkable so far. Five months later, he's back with what turns out to be contiguous spread to the medial cheek, moderately differentiated, now invading muscle and cartilage, no perineural invasion, but needing three stages to clear, and reconstruction required a flap plus a full-thickness graft for mucosal lining. Then — and this is the part that should raise your eyebrows — just ten weeks after that Mohs clearance, he's back again with a new nodule right under the flap. Recurrence, moderately differentiated, now tracking to the undersurface of the maxilla. That took five stages to clear, scraping bone, a defect approaching six by four centimeters, and they deliberately deferred definitive reconstruction for monitoring purposes. He stops upadacitinib around this point. And even after stopping, two months later, he develops yet another recurrence, this time poorly differentiated and found in the masseter muscle on imaging, PET-positive, requiring excision with otolaryngology and a buccal fat flap interpolation repair, followed by adjuvant radiotherapy.

So the trajectory here is a well-differentiated, unremarkable-looking primary that behaves like a completely different tumor on each subsequent recurrence — dedifferentiating and deepening with each round, recurring within ten weeks of clear Mohs margins, which the authors rightly flag as atypical. There's no formal discussion of limitations here because it's a case report — the honest caveat is simply that this is one patient with multiple confounders. He'd previously been on mycophenolate, which doesn't have strong evidence for increased nonmelanoma skin cancer risk, so that's probably not the driver. He's got prior BCCs, smoking history, fair skin, and age all stacked on top of whatever the JAK inhibitor is contributing, so causality can't be pinned down — the authors are careful to frame this as a "potential association" and a "warning," not a proven mechanism.

The practical takeaway, and this one is genuinely actionable: this reinforces exactly what's already sitting in the upadacitinib package insert — periodic skin exams for patients with risk factors for nonmelanoma skin cancer. For us specifically, the message is that in patients on JAK inhibitors who have any of these compounding risk factors — prior skin cancer, smoking, fair skin, age — a squamous cell carcinoma that recurs unusually fast or dedifferentiates on recurrence should prompt early multidisciplinary involvement sooner rather than later, meaning Mohs plus otolaryngology plus radiation oncology at the first sign of aggressive behavior, not after several rounds of surgery. This isn't practice-changing in the sense of altering how you treat primary SCC, but it should raise your index of suspicion and lower your threshold for tumor board discussion in this specific patient population.

Next, a retrospective cohort study: "Effect of Zinc Oxide Compression Dressing on Complications Post Lower Extremity Mohs Surgery." This one actually has a comparison group and real statistics, so let's walk through it properly.

The clinical problem is a familiar one to anyone doing Mohs below the knee — these wounds are notoriously complication-prone, and zinc oxide compression dressings, the Unna boot, are a staple for chronic venous wounds, but the evidence for using them prophylactically after Mohs surgery specifically is thin. The prior data is basically one small case series of ten patients with no complications, plus some traditional excision literature suggesting a dramatic complication reduction with Unna boots — zero percent versus about fourteen percent in one study. But that's excisional surgery, not Mohs, so the authors set out to fill that specific gap.

The design is a two-surgeon retrospective cohort from University of Rochester, essentially a natural experiment: one surgeon used zinc oxide dressings routinely on lower extremity Mohs cases, the other used standard wound care, so you get a built-in comparison group by surgeon preference rather than randomization. They identified 115 cases in each arm and used 1-to-1 greedy matching on BMI, immunosuppression status, diabetes, and smoking history to balance the groups before comparing outcomes.

Worth pausing on the statistics here because it's a nice teaching moment on matching. Greedy matching is a technique where, for each patient in your exposed group, the algorithm searches your comparison pool and grabs the closest available match on your chosen variables, pairing them off one at a time without going back to reoptimize globally — it's called "greedy" because it takes the best available match immediately rather than solving for the best overall pairing across the whole dataset. This matters here because with only two surgeons and non-randomized treatment assignment, patient selection could easily be confounded by exactly those variables — a surgeon might be more likely to reach for a compression dressing in a diabetic or immunosuppressed patient specifically because they're worried about wound healing. Matching on those variables upfront is the authors' attempt to strip out that confounding before ever comparing complication rates, which is a reasonable substitute for randomization when you're stuck working retrospectively. Chi-squared tests were then used to compare categorical outcomes like complication frequencies between groups, and Mann-Whitney tests handled the continuous variables like age and defect size — appropriate choices, though the authors don't spend time justifying them beyond convention.

Onto results. The two groups ended up well matched apart from a small age difference — the dressing group was about four years younger on average, which reached statistical significance but is not something you'd call clinically meaningful. Tumor types were mostly squamous and basal cell carcinoma, as you'd expect for this location. And the core finding: complication rates were essentially identical between groups. Dehiscence ran about five percent versus six percent, surgical site infection about seven percent versus four percent, neither difference statistically significant, and nobody in either group needed narcotic pain management. There was a trend toward less postoperative bleeding with the zinc oxide dressing — under one percent versus about four percent — but this didn't reach statistical significance, and critically, no patient in either group had bleeding requiring intervention, so even if that trend were real, it's not obviously changing anyone's clinical course. The one difference that was both statistically significant and unambiguous was postoperative visit frequency — patients with the zinc oxide dressing came back to clinic roughly twice as often within sixty days, about one and a half visits versus under one visit on average.

The authors' honest discussion here is refreshingly restrained — they don't oversell a null result. They note the increased visit burden is probably just from dressing changes, which might be reframed as a feature rather than a bug since it gives more chances to catch problems early, but they're explicit that with identical complication rates and no bleeding events requiring intervention in either arm, the clinical significance of the bleeding trend is limited. Limitations are the obvious ones for this design: single center, only two surgeons, retrospective, and the treatment assignment wasn't randomized — it was essentially surgeon-dependent, which even with matching leaves room for unmeasured confounding in surgical judgment.

Practical takeaway: this is not practice-changing in the sense of telling you to adopt or abandon zinc oxide dressings. What it does is validate that standard wound care is a perfectly reasonable default for lower extremity Mohs wounds — you're not exposing patients to excess complication risk by skipping the Unna boot. If you have a patient with a large defect, poor home wound care capacity, or venous insufficiency where the extra structural support and compression genuinely make physiologic sense, zinc oxide dressing remains a reasonable option, but the decision should be shared with the patient given the added visit burden and the essentially equivalent outcomes. Think of this as evidence-based permission to keep doing what you're already doing, rather than a mandate to change anything.

Third, and the most data-rich piece in this issue: "Real-World Use and Complications of Skin Substitutes in Mohs Micrographic Surgery: A Retrospective 10-Year Analysis." This is a retrospective cohort study using the TriNetX federated network, which pools electronic health record data from over seventy health care organizations nationwide.

The gap being addressed is straightforward — skin substitutes, meaning acellular dermal matrices, xenografts, and placental allografts, have become common adjuncts for closing large or complex Mohs defects, especially in older or comorbid patients who aren't candidates for extensive flap work, but there's essentially no large-scale, head-to-head safety data comparing these products against each other in the specific context of Mohs surgery.

Methodologically, the authors queried ten years of data, 2014 through 2024, using ICD-10, CPT, and the Q-codes that specifically identify individual skin substitute products, capturing cases where a skin substitute graft procedure was billed same-day as Mohs surgery — a reasonable proxy for immediate postoperative use. They split cases into two anatomic cohorts based on CPT code: one covering high-complexity areas like head, neck, hands, feet, genitalia, and sites involving cartilage, bone, tendon, or major vessels, and the other covering trunk and extremities. Within each cohort, they looked at the ten most commonly used products and tracked infection and bleeding as complications temporally associated with the index procedure. Why this design? The authors don't spend much time justifying it explicitly, but it's the obvious choice here — this is a big-network claims-based approach because getting head-to-head randomized data on ten different specialty wound products across a full decade would be essentially impossible to fund or enroll; a retrospective network query is really the only feasible way to generate comparative real-world signal at this scale. The tradeoff, which the authors are upfront about, is that these are temporally associated outcomes, not proven causal ones — there's no adjudication of whether the skin substitute itself caused the infection versus the underlying wound characteristics.

On the results side, in the complex-anatomy cohort, they captured over eleven hundred uses across the ten products. Integra, the acellular dermal regeneration template, was by far the most frequently used product and grew steadily across the decade. Xenograft use, things like MatriStem, picked up starting around 2018, Kerecis rose starting in 2022, while the placenta-derived products like Epifix and Grafix Prime actually peaked earlier, around 2020 to 2021, and have since declined in relative use. In the trunk and extremity cohort, overall volume was much lower — only around 250 uses total — and the product mix flipped, with Oasis Wound Matrix and Grafix Prime dominating instead of Integra.

Now the safety comparisons, which are the clinically actionable part. In the complex-anatomy cohort, PuraPly, the porcine-derived matrix, had the highest infection rate of any product compared against Kerecis, Cytal, MatriStem, Epifix, AlloDerm, and Integra — all of those comparisons were statistically significant. Apligraf, the full-thickness skin equivalent, also showed significantly higher infection rates than Epifix, Kerecis, MatriStem, AlloDerm, and Integra. Grafix Prime showed higher infection rates specifically compared to Kerecis and MatriStem. On the bleeding side, PuraPly again stood out with higher bleeding risk than Kerecis, AlloDerm, MatriStem, and Cytal, and Apligraf again showed higher bleeding risk than Epifix, Kerecis, MatriStem, AlloDerm, and Integra. In the trunk and extremity cohort, the numbers were simply too small — ten or fewer complications per product — to run reliable statistical comparisons at all, which is an important caveat since that's the cohort where Oasis and Grafix Prime were actually the dominant products.

Worth a beat on why the statistical comparisons themselves matter here: with pairwise product-to-product testing across ten products, you're running a large number of individual comparisons, and the authors report these as significant at a stringent threshold, which does lend some confidence that these aren't just noise — but the sample sizes underlying rarer products like Apligraf and PuraPly were still small in absolute terms, twenty-one and eighteen uses respectively, so even a couple of extra infections in a small denominator can generate a large relative-risk difference. That's a caution about precision, not about direction — Integra, Kerecis, and MatriStem consistently look like the safer performers on both bleeding and infection across multiple comparisons, and that pattern showing up repeatedly across several products does add confidence.

The authors' own discussion ties product selection patterns to real-world practicality — Integra's dominance likely reflects its long track record in burn and reconstructive surgery and its versatility for large or bone-exposed defects, porcine matrices tend to get used for smaller wounds, and amniotic or chorionic allografts get reserved for delicate sites like the eyelid, all filtered through cost and availability, since bovine and porcine products are cheap and widely stocked while placental allografts are considerably pricier. Limitations are squarely acknowledged: reliance on billing codes rather than clinical documentation, thin outcome data in the trunk and extremity cohort, no visibility into one-stage versus two-stage substitute placement, and the fundamental inability to establish causality from temporally associated claims data.

Practical takeaway here: this is genuinely useful comparative signal, not just interesting trivia, particularly if you're selecting among products for complex head and neck defects — knowing that PuraPly and Apligraf carry a real, statistically demonstrated excess of both infection and bleeding relative to Integra, Kerecis, and MatriStem is the kind of thing that could reasonably influence your product choice on your shelf today, especially for higher-risk defects where you want to minimize complication burden. I'd stop short of calling it fully practice-changing given the claims-data limitations and the causality caveat, but for a decision you're already making empirically based on cost and availability, this gives you actual comparative safety data to fold into that decision, particularly steering away from PuraPly and Apligraf when a lower-risk alternative is equally suitable.

Last for this episode, a technique article paired with an invited commentary: "The Bazooka Nasal Splint: Supporting the External Nasal Valve After Mohs Micrographic Surgery." Pure technique piece, so we'll walk through the steps and rationale rather than hunting for results or limitations that aren't there.

The clinical problem is external nasal valve support after Mohs defects involving the ala — anytime your defect includes the caudal septum, columella, soft tissue triangle, alar rim, or nasal sill, you're at risk of postoperative valve collapse or insufficiency, and while reconstruction methods for this area are well described, temporary splinting as an adjunct is underutilized in the Mohs literature despite being common in otolaryngology.

The technique itself: you take a silicone sheet, specifically the authors specify 1.02 millimeter thickness silastic, because thinner sheets are too flimsy to provide structural support. You cut it to a height that supports the ala without protruding more than a millimeter or two past the nostril, fold it into a tube with overlapping edges, clamp it with hemostats to hold shape, and suture the overlap closed with fine polypropylene, tucking the knot inside the tube so it doesn't irritate the mucosa. Then you fashion a separate oval "lily pad" piece from the remaining sheet, which serves as a bolster against the septum on the contralateral side. The whole assembly gets coated with mupirocin, the tube is inserted into the ipsilateral nasal vestibule, and a horizontal mattress suture on a PS-2 needle passes through the lily pad, through the cartilaginous septum — and this point is emphasized, specifically the cartilaginous septum, not the membranous septum — and into the splint, then back out through the splint and lily pad to complete the mattress stitch, taking care not to place the second pass too close to the first or the sheet will tear. It stays in place about a week and is removed by clipping the suture in clinic.

The stated rationale for reaching for this tool: it lets you skip cartilage grafting for structural ENV support in some cases, and for full-thickness nasal lining defects of small-to-medium size, the mucosa can granulate over the splint, potentially avoiding the need to surgically replace the lining altogether. It also lets the patient breathe through the ipsilateral nostril during the peak postoperative edema window, which is a meaningful patient-comfort win that a fully packed nostril doesn't offer.

Now, the invited commentary that runs alongside this is genuinely worth your attention because it pushes back constructively rather than just praising the technique. The commentator agrees nasal valve insufficiency is real but underappreciated after Mohs reconstruction, arising from bulky flaps, graft contraction, inadequate structural support, mucosal scarring, or dead space from lidocaine volume, hematoma, or postoperative fluid collections. But they raise specific safety concerns about this particular technique: suturing through the vascular septum in patients who are frequently on anticoagulants creates real bleeding or hematoma risk, potentially requiring unplanned visits for cauterization; suturing through an already-thinned or perforated septum — which many patients have without knowing it — risks worsening or creating a perforation; and the need for in-office suture removal is a burden for patients traveling long distances or for staff unfamiliar with the technique. The commentator also notes practical fall-through risks like dementia or sundowning patients tugging at the splint overnight. Their personal alternative is simpler and non-sutured: prepurchased silicone cone-shaped intranasal breathing aids, essentially anti-snoring dilators, sized appropriately and left in place for one to two weeks, with the tradeoff that they're less secure but avoid any septal suturing at all.

So the practical takeaway on this one: this is a genuinely useful addition to your reconstructive toolkit for ENV support after nasal ala defects, and it's interesting enough that it's worth trying — but it is not a slam-dunk replacement for existing splinting options, and the commentary's bleeding and perforation concerns are legitimate enough that you should think carefully about anticoagulation status and baseline septal integrity before threading a suture through someone's cartilaginous septum. For low-risk patients without anticoagulation or septal pathology, this is a reasonable technique to add. For patients on blood thinners or with any suspicion of septal thinning, the commentator's non-sutured cone alternative — or another non-perforating splint — may be the more prudent first choice.

That wraps our four articles for this September 2026 issue — a sobering JAK inhibitor case worth flagging in your immunosuppressed patients, reassuring null data on lower extremity wound dressings, real comparative safety signal on skin substitute selection, and a new nasal splinting technique to weigh against its own commentary's caution. Thanks for listening, and we'll see you next month.