Wukich DK, Joseph A, Ryan M, Ramirez C, Irrgang JJ.  Outcomes of ankle fractures in patients with uncomplicated versus complicated diabetes. Foot Ankle Int.  2011 Feb; 32(2): 120-30.     (PubMed ID: 21288410)

WHY did the authors undertake this study?
As I have quickly learned from working in North Philadelphia, not everyone who breaks their ankle is a great surgical candidate.  Dr. Wukich begins this article with an excellent and concise literature review detailing the potentially deleterious effects of diabetes on outcome following surgical correction of ankle fractures, and hypothesizes for this study that a group of patients with uncomplicated diabetes would experience fewer complications following ankle fracture ORIF compared to a group of patients with complicated diabetes.

HOW did they attempt to answer this question?
The primary outcome measure of this study was the development of a complication following ankle fracture ORIF.   Complication was defined as superficial infection, deep infection, nonunion, malunion, development of Charcot neuroarthropathy, requirement of amputation and/or the need for fixation revision or ankle fusion.

Diabetic patients were divided into two groups:  complicated (presence of end organ damage including peripheral neuropathy, nephropathy and/or peripheral arterial disease) versus uncomplicated (none of these associated conditions).  Patients were identified over a 5-year retrospective data collection period from a medical record review of two teaching hospitals.

WHAT were the specific results?
Statistically significant differences were found with respect to the overall complication rate (complicated diabetic patients were 3.8 times as likely to develop any complication), total non-infectious complication rate (complicated diabetic patients were 3.4 times as likely to develop a non-infectious complication) and requirement for revisional surgery (complicated diabetic patients were 5.0 times as likely to require a revisional surgery).  As with another of Wukich’s studies, patients with neuropathy were found to be more likely to develop an infectious complication.

HOW did the authors interpret these results?
From these results, the authors concluded  that patients with complicated diabetes carry an increased risk for the development of post-operative complication following ORIF of ankle fracture.

Langsetmo L, Hanley DA, Prior JC, Barr SI, Anastassiades T, Towheed T, Goltzman D, Morin S, Poliquin S, Kreiger N, CaMos Research Group.  Dietary patterns and incident low-trauma fracture in postmenopausal women and men aged > 50 y: a population-based cohort study.  Am J Clin Nutr. 2011 Jan; 93(1): 192-9..  2009 Sep; 109(3): 943-50. (Pubmed ID#: 21068350)

WHY did the authors undertake this study?
We know from previous studies that changes in dietary patterns may contribute to better bone health, and that better bone health may contribute to reducing the risk of fracture in older populations, but the aim of this study was to directly assess whether dietary patterns are related to the risk of fracture in a cohort of patients independent of other lifestyle variables.

HOW did they attempt to answer this question?
This is a retrospective cohort subgroup analysis from the larger Canadian Multicentre Osteoporosis Study (CaMos).  Dietary patterns of all participants were assessed with the use of a self-administered food-frequency questionnaire at the beginning of the study, and patients were followed for the development of a low-trauma fracture occurring before the scheduled 10th annual follow-up.

WHAT were the specific results?
“Nutrient dense” diets (i.e. good things like fruits, vegetables and whole grains) reduced the risk of fracture in both men and women compared to “energy dense” diets (i.e. bad things like soda, chips, meats and desserts).  The most significant findings were noted in older women.

HOW did the authors interpret these results?
From these results, the authors concluded the benefits of a healthy diet probably extend into bone health and fracture risk.

I have to admit that I somewhat balked the first time I scanned through the Wukich et al diabetic ankle fracture study.  After reading the title, I thought it sounded like an interesting and appropriate clinical question, and immediately looked to the tables to see how the authors defined “complicated diabetes” versus “uncomplicated diabetes”.  As with most well-written comparative clinical studies, I was able to flip to Table 1 and easily identify the specific patient characteristics of the two patient cohorts. And also like most well-written comparative clinical studies, Table 1 was set up in a pretty standardized way:

The reason that I initially balked was because I was surprised to see in this table that there were no statistically significant differences in hemoglobin A1c values between the two groups. In fact, they were essentially the same (7.0+/- 0.61 versus 7.0+/-0.87; I calculated a quick p-value for these and found that it was essentially 1.0). I had assumed that the authors would divide the “complicated” versus “uncomplicated” diabetic patient cohorts based on the HbA1c values. In fact, the authors defined “complicated diabetes” as the presence of end-organ damage including peripheral neuropathy, peripheral arterial disease, and/or nephropathy (a perfectly appropriate thing to do). The presence of these co-morbidities was statistically significant between the two cohorts on Table 1. So my initial assumption about the establishment of population cohorts in this study was wrong.

I think this brings up a good point and a potential criticism of retrospective study designs in that authors can essentially establish the cohorts based upon whichever criteria they want after-the-fact. Over the last couple years Dr. Wukich has published some very sound and very interesting studies, all asking similar questions in slightly different patient cohorts.  In 2009, he published on “early complications” (the variable) in patients “with diabetes undergoing pilon fracture ORIF” (cohort #1) and “without diabetes undergoing pilon fracture ORIF” (cohort #2). In 2010, he published on “postoperative infection rates” (the variable) in patients “with diabetes undergoing foot and ankle surgery” (cohort #1) and “without diabetes undergoing foot and ankle surgery” (cohort #2). And here, he published on “outcomes” (the variable) in patients with “complicated diabetes undergoing ankle fracture ORIF” (cohort #1) and “uncomplicated diabetes undergoing ankle fracture ORIF” (cohort #2). The patients and interventions were all relatively similar, but by dividing up the patient groups in different ways, he was able to ask slightly different questions. 

From Dr. Wukich’s extensive bank of surgical patients, we can essentially divide them up into as many different ways as we can think of: diabetics versus non-diabetics; complicated diabetics versus uncomplicated diabetes (based on end-organ damage); controlled diabetics versus uncontrolled diabetics (based on HbA1c values); neuropathic diabetics versus sensate diabetics (based on SWMF testing); and on and on and on.

These type of studies are characterized as Level II in the “Prognostic Study” column in our levels of clinical evidence because of their retrospective nature, and part of their critical analysis involves ensuring that the specific patient characteristic that separates the groups is clinically valid.