Risk of complications and poorer postoperative outcomes in obese and diabetic patients following upper limb arthroplasty: a systematic review and meta-analysis

Abstract

The impact of obesity and diabetes mellitus on patient outcomes following upper limb arthroplasty is contentious. With increasing demand for joint arthroplasty, risk factors that predispose patients to greater complications and poorer outcomes must be thoroughly investigated. The objective of this review was to synthesise the best available evidence investigating the influence of obesity or diabetes mellitus on complications and/or poorer postoperative outcomes following total shoulder (TSA), reverse total shoulder (RTSA) and total elbow arthroplasty (TEA). Electronic databases (PubMed, CINAHL, and Embase) and grey literature were searched for studies that evaluated the influence of obesity (Body Mass Index [BMI] ≥ 30 kg/m²) or diabetes mellitus on arthroplasty outcomes. Two independent reviewers assessed the methodological validity of eligible studies and data was pooled in statistical meta-analysis where appropriate (RevMan 5.3; Cochrane Collaboration). The review was prospectively registered on PROSPERO (CRD42016053299). Twenty-one studies (20 cohort studies and one case-control) were included. The majority of studies considered TSA and/or RTSA populations, while four studies evaluated TEA patients. Obesity was found to significantly increase operative duration, with obese TSA/RTSA patients experiencing operations 10.00 minutes longer (95% CI [6.31, 13.69]) than patients with a BMI in the normal range, which increased to 12.48 minutes (95% CI [8.40, 16.55]) in patients with a BMI ≥ 35.0. Furthermore, obese and morbidly obese patients had 3.92 (95% CI 3.59, 4.28) to 5.46 (95% [CI 4.91, 6.07]) times greater odds of venous thromboembolism compared to their non-obese counterparts. Similarly, odds of infection increased with increasing BMI, from 2.37 (95% CI [1.65, 3.41]) times in obese, to greater than five times (95% CI [4.70, 5.39]) in morbidly obese. Obesity also increased the odds of revision (OR 1.52; 95% CI [1.43, 1.61]), dislocation (OR = 2.51; 95% CI [2.35, 2.69]) and fracture (OR = 1.94; 95% CI [1.79, 2.10]) in TSA, RTSA and TEA patients, however had no influence on the odds of urinary tract infection (OR = 0.88; 95% CI [0.48, 1.61], length of stay (MD = 0.15; 95% CI [-0.28, 0.58]), unscheduled return to theatre (OR = 0.74; 95% CI [0.44, 1.24]) or mortality (OR = 1.79; 95% CI [0.79, 4.03]). Nonetheless, morbid obesity made a small, yet significant, difference on mean length of stay (MD = 0.28; 95% CI [0.14, 0.43]). Evidence examining the effect of obesity on blood transfusion was inconclusive, while minimal evidence was available on pneumonia and quality of life. Diabetic TSA, RTSA and TEA patients had 2.93 (95% CI [1.97, 4.35]) times greater odds of mortality as an inpatient. Furthermore, diabetes mellitus was found to significantly affect odds of blood transfusion (OR = 1.49; 95% CI [1.41, 1.57]) and pneumonia (OR = 1.38; 95% CI [1.14, 1.67]), however had no effect on the odds of pulmonary embolism (OR = 1.17; 95% CI [0.94, 1.44]). The evidence for greater risk of blood transfusion in diabetic patients is a concern given the higher odds of further complications observed in transfused patients. There was also limited evidence on unscheduled return to theatre and urinary tract infection. No evidence was found examining the impact of diabetes mellitus on operative duration, dislocation, fracture, pain, function, quality of life, and revision. Inferences are limited for a number of the outcomes due to methodological shortcomings and confounders. Operative duration was inconsistently defined, and prophylactic regimes for infection and venous thromboembolism were not standardised and varied, across the included studies. The literature suggests that patient factors such as age and gender influence outcomes including revision, infection and fracture, and that surgical factors may impact the incidence of dislocation. A major limitation of studies investigating diabetes mellitus was that they reported data grouped by diabetes mellitus diagnosis without reporting the criteria used for diagnosis, or the level of glycaemic control at time of surgery. A further inherent limitation is the low level of evidence of observational study designs commonly used in orthopaedic research. Surgeons are advised to consider the additional risks associated with obesity and diabetes mellitus when determining optimal treatment options for upper limb arthroplasty patients.