Diagnosis and management of fracture-related infections in a low-income countrya prospective study comparing current practice to international consensus guidelines

What happens when the international playbook for treating bone infections meets a place that cannot afford the playbook? That question sounds rhetorical. Sliepen and colleagues turned it into a study, and the answer is both more specific and more uncomfortable than you might expect. Fracture-related infection, or FRI, is an infection at the site of a healing bone fracture. In practical terms, picture a femur or tibia with metal hardware inside it that simply will not unite. The bone isn't healing because microbes have taken hold, and those microbes are causing damage that no amount of waiting will reverse. Sliepen and colleagues describe it as "a complex and often difficult-to-treat disease entity" that burdens patients, physicians, and healthcare systems alike. International consensus groups have responded by building formal diagnostic and treatment frameworks. These frameworks separate signs of infection into stronger, confirmatory findings and weaker, suggestive ones, and specify that even a stalled fracture, or non-union, should be treated as a potential infection until proven otherwise. Those frameworks also define the surgical options. The first is DAIR, which stands for debridement, antimicrobial therapy, and implant retention. This option is used when the fracture is still healing and the hardware can remain in place. The second option is implant removal or exchange combined with debridement. If the fracture hasn't healed yet, that exchange can happen in one operation or two. A single-stage procedure removes infected hardware and inserts new definitive fixation in the same sitting. A two-stage approach takes an interim step, which involves debridement, temporary stabilization, and antibiotics, before going back in for definitive fixation. Each strategy has its place, and knowing which to use, and whether it was actually used correctly, requires good microbiological data and a validated definition of infection. Sliepen and colleagues set out to assess whether that standard is being met. They conducted a systematic review guided by the PRISMA criteria, registered in PROSPERO. They screened more than twenty-one thousand records and landed on thirty-five studies covering nine hundred eighty-five patients with long-bone FRIs of the femur, tibia, and humerus. Twenty-five studies were retrospective, ten were prospective, and one was randomized. Follow-up was inconsistent: twelve studies had at least twelve months, eleven had shorter follow-up, and twelve did not specify a minimum at all. Methodological quality was scored using the McMaster critical appraisal tool; single-stage studies averaged sixty-two percent, and two-stage studies averaged sixty-eight percent. Neither number signals strong evidence. The diagnostic picture is where the gap between guidelines and reality opens widest. Only seven of the thirty-five studies used a validated definition of infection. The rest relied on vague language such as "clinical evidence of infection" or "presence of active infection" — phrases that tell you almost nothing about what the clinician actually saw or measured. Only five studies included FRIs diagnosed within six months of primary fixation, leaving early infections almost entirely absent from the literature. Microbiological sampling was similarly inconsistent. Tissue cultures are the centerpiece of the consensus diagnostic pathway. You need to know which organism you're treating before you can treat it effectively. However, many studies did not describe their sampling protocols in enough detail to evaluate adherence. The consequence of skipping that step showed up clearly in data from presumed aseptic non-unions — fractures that looked clean but weren't. Two studies, by Amorosa and colleagues and Hackl and colleagues, reported that when intraoperative cultures came back unexpectedly positive, the rate of additional surgeries jumped to twenty-eight and forty-four percent, respectively. In the culture-negative patients from those same studies, additional surgery rates were six point four and eighteen percent. Wagner and colleagues found a similar pattern: twenty-two percent of culture-positive non-unions needed another operation, compared to six percent of culture-negative ones. The message is direct — unrecognized infection leads to worse outcomes, and you can only recognize it if you look. On the treatment side, the comparison between single-stage and two-stage procedures produced numbers that look similar on the surface but carry a lot of uncertainty underneath. Of the nine hundred eighty-five patients, eight hundred forty-five underwent single-stage revision, and one hundred forty underwent two-stage. Mean bone healing was eighty percent for single-stage and seventy-seven percent for two-stage. Infection eradication was eighty-seven percent and eighty-one percent, respectively. Additional procedures were needed in roughly twenty-eight percent of single-stage patients and twenty-five percent of two-stage patients. Those figures are close enough that Sliepen and colleagues conclude there is simply insufficient evidence to recommend one strategy over the other for unhealed long-bone FRIs without large bone defects. What the numbers obscure is how much variation existed within each category. Many centers used custom-made polymethyl methacrylate, or PMMA, coated implants, which is a cement-based coating that can carry local antibiotics. Twenty-three of the thirty-five studies used these devices; twelve used PMMA-coated non-conventional cores like K-wires, guide wires, and Ilizarov rods. The non-conventional coated implants, which lack locking mechanisms, showed lower mean bone healing — seventy-three percent compared to eighty-four percent for coated conventional implants in single-stage series. Many teams had to supplement them with a brace or cast just to maintain stability. Complications from PMMA were described as "rather common": cement debonding, nail bending or breakage, and extraction problems. A separate review put antibiotic-cement coating complication rates between five and thirty percent. Antimicrobial practice was the most heterogeneous element of all. Local antibiotics were used with varying agents, dosages, and no standardized protocols. Some centers started systemic intravenous antibiotics days before surgery; one study reported using no intravenous antibiotics at all because local antibiotics were deemed sufficient. Only four of the twelve studies using conventional uncoated implants even described local antimicrobial use. Sliepen and colleagues state plainly that this heterogeneity made it impossible to assess the impact of local antimicrobials, and they raise concerns about sub-therapeutic elution and resistance with PMMA-based delivery. Some of this divergence from consensus reflects genuine resource constraints. The authors document situations where a two-stage approach was chosen specifically because a plastic and reconstructive surgeon was not available for soft-tissue coverage. That's not poor decision-making — it's triage. But other divergences are measurement problems. If you're not using a validated infection definition, you can't compare your outcomes to anyone else's. If you're not taking tissue cultures, you're not following the diagnostic pathway, and you may be missing infections that will surface later as complications. That distinction between resource constraints and measurement constraints is the study's most useful contribution. Sliepen and colleagues found mean bone healing rates of eighty and seventy-seven percent across the two surgical strategies, and infection eradication rates of eighty-seven and eighty-one percent. But they are explicit that substantive heterogeneity and methodological weakness make those numbers hard to trust as guides for practice. Their recommendations are concrete: future studies should use validated definitions of infection and outcomes, standardize how additional procedures are counted, ensure at least one year of follow-up, and account for differences in patient health, fracture type, and soft-tissue status. They also point to a recently proposed FRI classification system as a tool that could help standardize how cases are categorized before treatment decisions are made. The gap between best-practice guidelines and actual clinical practice is not a single gap. It has at least two layers. One is the resource layer — what implants, specialists, and laboratory capacity are available. The other is the measurement layer — whether the field is even asking the same questions in the same language. Sliepen and colleagues show that on the measurement side, thirty-five studies, nearly one thousand patients, and more than two decades of published literature have still not produced an evidence base strong enough to make a firm call on one of the most fundamental surgical questions in fracture care. That finding should be uncomfortable. It is also, precisely because it is documented, a place to start. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.