Trends in Comorbidities and Complications Among Patients Undergoing Hip Fracture Repair

Janis Bekeris, MD; Lauren A. Wilson, MPH; Dace Bekere, MD; Jiabin Liu, MD, PhD; Jashvant Poeran, MD, PhD; Nicole Zubizarreta, MPH; Megan Fiasconaro, MS; Stavros G. Memtsoudis, MD, PhD, MBA


Anesth Analg. 2021;132(2):475-484. 

In This Article


In this retrospective cohort study of N = 507,274 patients undergoing hip fracture repair surgery, we observed a number of noteworthy trends. Between 2006 and 2016, overall comorbidity burden increased among patients undergoing hip fracture repair surgery. This coincided with an increase in postoperative acute renal failure and decline in postoperative mortality, pneumonia, hemorrhage/hematoma, and venous thromboembolism. Moreover, use of neuraxial anesthesia decreased over time.

In this study, most patients suffering from hip fractures were elderly and of female sex. These findings highlight the importance of osteoporosis prophylaxis among the elderly patient population, particularly postmenopausal women, with limited agility in combination with gradually decreasing bone quality.[17–20] Interestingly, while the population demographic profile stayed relatively unchanged, the rate of complex multilocated fractures significantly decreased over time, which may imply positive effects of establishing and encouraging respective prophylactic practice. From a procedural viewpoint, we saw almost twice as many patients receiving THA in 2016 when compared to 2006, while both internal fixation and hemiarthroplasty still remained the preferred treatment choice. The increase in THA reported here is in line with the findings by Miller et al.[21] Surgical volume has been suggested to influence postoperative outcomes in some surgical patient populations, including trauma patients, with high-volume centers performing better. However, more recent information regarding hip fracture repair remains equivocal,[22,23] and more robust evidence has been reported to be related to surgeon volume.[22,24] In our study, we observed a significant transition over time, where more patients were treated in small hospitals rather than medium- and large-sized health care centers. However, it is not possible to distinguish if this is merely a trend in the hospital size composition of Premier or if these procedures are actually being performed with greater frequency at smaller hospitals.

Throughout the study period, we saw a substantial increase in comorbidity burden. Among the observed preexisting comorbidities, the presence of obstructive sleep apnea increased almost 3-fold during the observation period. While this increase (and others) may be in part attributable to increased awareness, screening, and diagnosis,[25–28] the changes occurred rather gradually over time. Similarly and simultaneously, the prevalence of obesity, a major risk factor for obstructive sleep apnea,[25] more than doubled. This finding and the fact that hypertension prevalence also rose suggest that sleep apnea may indeed have become more frequent as obesity and metabolic syndrome become more prevalent in the United States.

Despite the increase in comorbidity burden, trends in the majority of the postoperative complications analyzed either remained constant or declined. There were substantial downward trends for acute myocardial infarction, hemorrhage/hematoma, pneumonia, and venous thromboembolism in the postoperative period. This suggests that measures to treat cardiovascular disease and minimize postoperative bleeding and infection may prove effective in preventing these potentially catastrophic complications. In line with this reduction in potentially life-threatening complications, we observed a significant decline in postoperative mortality rate throughout the study period. Substantial mortality reduction in various patient cohorts has been achieved with improvements in perioperative care. Pedersen et al[29] report a 40% mortality reduction after hip replacement between 1980 and 2014 in Denmark. However, further mortality rate reductions may be difficult to achieve because other precursors like frailty and cognitive decline within this aging cohort suggest a strong association to both fracture and subsequent death.[30] Further study is necessary to verify what is driving these declining trends and advance these improvements in perioperative care.

In addition, although the rate of blood transfusions declined throughout the study period, this change was not significant. While there is ample evidence that tranexamic acid reduces blood transfusions in hip fracture surgery, studies on optimal dosage and timing are limited.[31] The lack of information regarding the ideal administration of antifibrinolytics may be why blood transfusion rates have remained fairly high among hip fracture repair patients.

It is of interest that we found significantly decreased neuraxial anesthesia use over time. The reduction in the use of spinal anesthesia is noteworthy and somewhat surprising. Although the literature on the effect of anesthesia type in this patient population is equivocal at this point, numerous population-based studies suggest either better outcomes with neuraxial approaches or no impact.[8,32,33] It is, therefore, not surprising that clinical trials are currently underway to identify the impact of anesthetic type on outcomes in patients with hip fracture.[11,34] While the reasons for the reduction in the use of neuraxial anesthesia over time have to remain speculative using our data, it is possible that the increasing use of potent anticoagulants in this patient population and a push to earlier surgery may present anesthesiologists with fewer opportunities to utilize a neuraxial technique.

Our study has a number of limitations. First is the retrospective study design that does not allow us to determine the relationship between cause and effect. However, this study exclusively focuses on descriptive analyses. Another limitation is the reliance on ICD-9 and procedural codes to identify and to define the study variables that might occasionally under- or overcapture the occurrence of individual illnesses or adverse events. In addition, because we used a database covering a broad spectrum of hospitals, some interinstitutional differences in reporting and coding cannot be excluded with certainty. Potential biases arise from the possibility of billable diagnoses being reported more frequently than, for example, chronic illnesses.[35] Nonetheless, a rigorous effort is made by Premier to review and validate data for research purposes. While Premier consists of data from approximately 20% to 25% of all hospital discharges in the United States,[15] there are some patient populations that are not captured within the database (eg, patients treated at Veterans Affairs hospitals). Unfortunately, our findings cannot be extrapolated to these unrepresented patient populations. Last, the large sample size utilized in this study resulted in a high level of power, which we attempted to control for using a stringent statistical significance cutoff. A preliminary simplified power analysis demonstrated that our cohort was sufficiently powered to detect differences between 2006 and 2016. We have reported the difference that we had 90% power to detect for each outcome based on their prevalence in 2006 and reported these values in Supplemental Digital Content 2, Appendix 2, These differences are all fairly small due to our large sample size; subsequently, some of the significant trends observed in demographic characteristics were not clinically meaningful (eg, the 2% change in sex composition). However, in the context of trends in comorbidities and complications, our statistically significant findings were clinically significant as well.

In conclusion, we observed moderate changes in demographics among patients undergoing surgery for hip fracture between 2006 and 2016. Comorbidity burden increased over time, while the majority of complication rates either declined or remained largely unchanged. Along with other changes, anesthetic practice changed with neuraxial anesthesia being used less frequently. These results can be utilized by clinicians, researchers, administrators, and policy makers to assess and address emerging challenges and allocate resources accordingly.