Acetabular dysplasia is known to be a cause of early-onset hip osteoarthritis, as well as hip pain and disability in children and young adults 1,2. Earlier diagnosis and management of hip dysplasia is optimal, as younger patients benefit from greater remodeling potential 3,4. Pelvic osteotomies are commonly performed procedures to address hip dysplasia in pediatric patients. The long-term goal of the pelvic osteotomy is to prevent early-onset arthritis and obviate the need for arthroplasty at a young age. Multiple osteotomy techniques have been described, and the choice of procedure largely depends on the age of the patient, the severity of the disease, and the overall morphology of the hip joint 5.
While pelvic osteotomies in pediatric patients have been shown to have relatively low overall morbidity and mortality, there is limited data on the risk factors for perioperative morbidity. Existing studies have been limited by sample size, narrow outcome measures, and low generalizability 6–9. Further information is needed to better define the perioperative safety of these procedures.
Childhood obesity is one potential risk factor that has been receiving increased attention in recent years, as it is a growing public health concern in the USA 10,11. While there have been significant efforts in recent years to decrease the rate of childhood obesity, most of these have had a minimal effect 12. With the increased rates of childhood obesity in the USA, it is important to characterize the associated risks for patients undergoing osteotomies about the hip to facilitate perioperative discussions and risk stratification. The purpose of the present study is, therefore, to characterize differences in adverse events, operative time, length of stay, and readmission following pelvic osteotomy with or without femoral osteotomy for obese and nonobese patients using a large, prospectively-collected national pediatric surgical registry.
Patients and methods
A retrospective cohort study was carried out using the American College of Surgeons’ National Surgical Quality Improvement Program (NSQIP) Pediatric database, which captures data from participating children’s hospitals across the USA 13,14. In the NSQIP Pediatric database, 129 patient variables are prospectively collected from operative reports, medical records, and patient interviews by trained surgical clinical reviewers 13. Clinical data are subjected to internal audits to ensure inter-rater reliability above 2% for all variables. Patients are followed for the entire 30-day postoperative period, including after hospital discharge 14.
Following Institutional Review Board approval, the NSQIP Pediatric database from 2012 through 2014 was queried to identify patients under the age of 18 years who had the following Current Procedural Terminology codes: 27 146 (pelvic osteotomy), 27 147 (pelvic osteotomy with open reduction of hip), 27 151 (pelvic and femoral osteotomy), or 27 156 (pelvic and femoral osteotomy with open reduction of the hip).
The NSQIP Pediatric database collects numerous demographic and comorbidity variables for each patient, including age, sex, height, and weight, along with other comorbidity variables 14. Each patient’s height and weight were used to calculate BMI. BMI was categorized according to the Center for Disease Control and Prevention (CDC) BMI-for-age charts 15. As per the CDC guidelines, patients with BMI greater than or equal to the 95th percentile BMI-for-age were categorized as obese 11,12,16. History of pulmonary disease was divided into the history of asthma (within 1 year of surgery) or history of the nonasthma pulmonary disease (which included structural pulmonary abnormalities, history of bronchopulmonary dysplasia, or chronic lung disease). Information on impaired cognitive status and the presence of a neuromuscular disorder were also available in the dataset.
The NSQIP Pediatric database tracks patients for the occurrence of individual adverse events that occurred within the first 30 postoperative days. The following adverse events are recorded: death, coma, on ventilator for more than 48 h, stroke, thromboembolic events (pulmonary embolism or deep vein thrombosis), cardiac arrest requiring cardiopulmonary resuscitation, renal failure, sepsis, unplanned return to the operating room, wound dehiscence, surgical site infection (SSI, including superficial SSI, deep SSI, or organ/space infection), graft/prosthesis/flap failure, urinary tract infection, pneumonia, progressive renal insufficiency, and peripheral nerve injury. ‘Any adverse event’ was defined as a binary variable that was positive when any of the above adverse events occurred in the postoperative period.
The occurrence of a blood transfusion from the beginning of the procedure until 3 days postoperatively is also recorded in the database. Transfusion was not counted as a true adverse event, as the incidence of blood transfusion following femoral/pelvic osteotomy is high enough that its inclusion would skew outcomes.
Readmission was defined as an unplanned admission to any facility after the initial postoperative discharge within 30 days of the procedure.
Operative time was defined as the time in minutes from incision to the end of wound closure. Length of stay was defined as the number of calendar days a patient was in the hospital following the procedure.
Statistical analyses were carried out using Stata, version 13.1 (StataCorp, LP, College Station, Texas, USA). All demographic and comorbidity variables were compared between primary and revision procedure groups using χ 2 analysis. Continuous perioperative outcomes (operative time and length of stay) were compared using bivariate and multivariate linear regressions. Binary perioperative outcomes (adverse events, blood transfusion, and readmission) were compared using bivariate and multivariate Poisson regression with robust error variance. Poisson regression with robust error variance was used as an alternative to logistic regression in this analysis, as it has been well-established that it enables a more accurate statistical comparison of risks when the outcomes tested are not rare 17,18. Multivariate regression was used in this study to control for baseline patient characteristics, including age and procedure type, in all final analyses.
A total of 996 patients were identified (Table 1). The average age was 8.0±4.6 years (mean±SD). Fifteen percent of patients had BMI above the 95th percentile-for-age per CDC guidelines. Significant differences were identified between groups, with older patients having greater BMI percentile-for-age (P=0.002). There were more isolated pelvic osteotomies and fewer pelvic and femoral osteotomies performed in patients with higher BMI-for-age, compared with patients with lower BMI-for-age. There was a decreased rate of impaired cognitive status in patients with BMI-for-age above the 95th percentile.
A total of 5.72% of patients had any adverse event, with a 3.61% rate of severe adverse events, and a 3.21% rate of readmission (Table 2). On multivariate analysis, there was an increased risk of blood transfusion (relative risk: 1.4, P=0.007) and readmission (relative risk: 2.3, P=0.032) within 30 days for patients with BMI-for-age above the 95th percentile (Table 3). Obesity was not associated with the development of surgical site infection on multivariate analysis.
Using multivariate linear regressions comparing operative time and postoperative length of stay by BMI, no significant differences were found for either operative time or length of stay (Table 4). Reasons for readmission within 30 days following the procedure were also available for the majority of readmitted patients (Table 5). The most common reason was surgical site infection [six (16%) patients].
A total of 14 (2.60%) patients had a surgical site infection following the procedure. Half (50%) of these patients (seven patients total) were readmitted to the hospital within 30 days following the initial postoperative discharge. Among these seven readmitted patients, the reason for readmission was listed as infection for five patients, wound disruption for one patient, and no specific readmission reason was given for the remaining patient.
Pelvic and femoral osteotomies in pediatric patients have had success in treating hip dysplasia; however, the risk factors for increased perioperative morbidity for these procedures are poorly understood. Childhood obesity is a growing problem in the USA and has been shown to be a predictor of increased complication rates in pediatric orthopedic surgery 10. The purpose of the present study was to characterize differences in adverse events, operative time, length of stay, and readmission following pelvic osteotomy with or without femoral osteotomy for obese and nonobese patients. The present study identified obesity as an independent risk factor for blood transfusion and readmission following pelvic osteotomy with or without femoral osteotomy; however, obesity was not associated with increases in operative time or postoperative length of stay.
The rates of long-term complications, such as osteonecrosis and need for further surgical intervention, have been well studied; however, there is little literature assessing short-term adverse events after pelvic osteotomies in children 6. In a study by Lalonde et al. 7, the authors described results of osteotomies for correction of residual dysplasia in two pediatric age groups and found no immediate perioperative complications in either group. However, in a study of adult patients who underwent Bernese periacetabular osteotomies (PAO), Novais et al. 8 identified obesity as a risk factor for postoperative complications. The authors found that obese patients developed a major complication in 22% of cases, as compared with 3% in nonobese patients. In their study, the authors did not restrict their complications to the immediate postoperative period. In a small cohort study examining outcomes following PAO in adolescent patients, the authors found that obesity was the only variable that was significantly associated with postoperative complication rates 9. While there was no difference in radiographic surgical correction between the obese and nonobese cohorts, obese patients were 10 times more likely to develop a complication. The most common complications noted were superficial hematoma and infection. While these studies implicate obesity as a risk factor for postoperative complications, they are limited by low numbers of patients and may not be generalizable to all pelvic osteotomies.
The NSQIP Pediatric database consists of multicenter data chronicling individual adverse events, as well as blood transfusion and readmission rates, that occurred within the first 30 postoperative days. A previous study using the NSQIP Pediatric database examined short-term adverse events in the most commonly performed inpatient pediatric procedures. The authors found that overall, adverse events were relatively low (~7%) following pelvic osteotomies; however, the rate of blood transfusion is relatively high (~35%) as compared with other nonspine pediatric procedures. The authors also noted that for any procedure, obese patients (as defined by a BMI of ≥95th percentile) were at greater risk of perioperative complications, including infection, blood transfusions, and readmission 19. In our study, we found that obesity was an independent risk factor for blood transfusions and 30-day readmission, but not overall adverse events.
The finding that obesity was independently associated with increased risk of blood transfusion raises the concern that this relationship is because of possible unmeasured confounding factors such as preoperative blood volume or amount of surgical exposure. While the NSQIP Pediatric database does collect preoperative hematocrit (postoperative hematocrit is not collected), there is, unfortunately, a large amount of missing data for this variable, as this value is only recorded at select participating pediatric NSQIP centers. In the present patient sample, only 52.8% of patients had a preoperative hematocrit value. Although including hematocrit in analyses was initially considered, we believed that the high rate of missing data would likely significantly skew the results. This has been previously shown to occur when using this variable for the adult NSQIP 20. However, a post-hoc exploratory analysis was performed, using bivariate and multivariate Poisson regression with robust error variance, to assess a potential association between blood transfusion and preoperative hematocrit in those patients that had this information available to examine a possible physiologic explanation for the greater risk of blood transfusion in obese patients. Although this analysis was limited to the reasons given above, we found no statistically significant association between blood transfusion and preoperative hematocrit (P=0.080). The area of surgical exposure was controlled for in these analyses by the inclusion of the procedure type variable, which differentiates procedures with isolated pelvic osteotomy versus those that also included femoral osteotomy, which typically requires additional exposure. Finer details about the amount of surgical exposure are not available. The greater risk of transfusion in these obese pediatric patients still may be related to a larger surgical exposure or blood volume rather than a physiologic difference; however, the surgical registry used in the present study is not able to explore this further. The authors believe that the benefits of using prospectively collected data with complete 30-day follow-up from a large, national sample of patients outweigh the disadvantages from a lack of fine detail for certain variables. Further studies that are specifically designed to assess these outcomes in finer detail are warranted.
While risks associated with blood transfusions such as blood-borne infections have greatly improved over time, serious adverse events, such as transfusion-related acute lung injury and acute transfusion reactions, may still occur 21. An annual report of transfusion complications found that in pediatric patients, the majority (58%) of adverse reactions occurred secondary to human error, often related to unfamiliarity with special requirements in this population 22. In addition, there is a concern in the literature that immunomodulation following blood transfusions may lead to an increased risk of infection, postoperatively 23. Thus, it is important to recognize and potentially modulate factors that increase patients’ risk of blood transfusion, and based on the results of this study, obesity may be one of those factors.
Obesity was not found to be an independent risk factor for surgical site infection in the present study, yet obesity was an independent risk factor for 30-day readmission, and surgical site infection was the most common reason for readmission (16% of total readmissions). Not only is short-term readmission frustrating for patients, families, and providers, but readmission rates are now being used as an outcome to measure healthcare quality at many centers. Obesity may make outpatient postoperative care more difficult by complicating hygiene, therapy, and pain medication dosage, among other factors, leading to possible readmission. Despite the lack of a direct association with obesity in the present study, infection following pelvic osteotomy warrants further exploration as the most common reason for readmission. Obesity has been previously linked with infection following PAO by Novais et al. 9; however, their study used a sample of patients aged older than 18 years and used adult cutoffs for BMI, which limits direct comparison. Further detail regarding infection in the NSQIP dataset is limited, and future studies should investigate the role of antibiotics, different causative organisms, and other possible risk factors for surgical site infection following pediatric pelvic osteotomy.
There are several limitations to our study secondary to the characteristics of the NSQIP database. It is important to note that the database only reports adverse events that occur within the 30-day post-operative period, so late-onset adverse events are not captured in this data set. Thus, the rate of avascular necrosis of the hip, a common and important complication following treatment of hip dysplasia, is not described in our study. Furthermore, orthopedic-specific outcomes, such as pain, radiographic results, and functional scores, are not reported by the database. In addition, the database does not distinguish between different osteotomy techniques, so our group of patients represents a relatively heterogeneous cohort. Furthermore, we included patients who underwent femoral in addition to pelvic osteotomy to maintain adequate power for analyses. Similarly, patients with neuromuscular diseases and those with impaired cognitive status were also included to appropriately power analyses. Procedure type, neuromuscular disease, and cognitive status were all included as variables in the multivariate regression, which controlled for the inclusion of these patients. Another limitation is that data of potential interest related to infections, such as the isolated organisms and/or antibiotics used, were not collected in the dataset and are, therefore, unavailable for analysis. Future studies should attempt to better define the infection profile in this patient population. Finally, use of the NSQIP database precludes us from accounting for surgeon experience, which may have an overall effect on surgical outcomes. Despite these limitations, we believe that the large patient numbers and high-quality data collection process employed by the NSQIP contributes to the strength of our study. Furthermore, the cohort of NSQIP participating centers is comprised of a variety of community and tertiary referral centers, which improves the generalizability of our results. To our knowledge, this is the largest study to characterize the risks of perioperative complications associated with obesity following pelvic osteotomies in children.
Using a large, prospectively collected cohort of pediatric orthopedic patients, this study has identified obesity as an independent risk factor for blood transfusion and readmission following pelvic osteotomy with or without femoral osteotomy. Our data provide information regarding the risks of pelvic osteotomy in obese patients and can facilitate patient counseling by physicians and informed decision-making by patients and families when planning for surgical treatment of hip dysplasia.
Conflicts of interest
There are no conflicts of interest.
1. Wedge JH, Wasylenko MJ. The natural history of congenital disease of the hip. J Bone Joint Surg Br 1979; 61-b:334–338.
2. Weinstein SL. Natural history of congenital hip dislocation (CDH) and hip dysplasia
. Clin Orthop Relat Res 1987; 225:62–76.
3. Harris NH. Acetabular growth potential in congenital dislocation of the hip and some factors upon which it may depend. Clin Orthop Relat Res 1976; 119:99–106.
4. Salter RB, Dubos JP. The first fifteen year’s personal experience with innominate osteotomy in the treatment of congenital dislocation and subluxation of the hip. Clin Orthop Relat Res 1974; 98:72–103.
5. Gillingham BL, Sanchez AA, Wenger DR. Pelvic osteotomies for the treatment of hip dysplasia
in children and young adults. J Am Acad Orthop Surg 1999; 7:325–337.
6. Vitale MG, Skaggs DL. Developmental dysplasia of the hip from six months to four years of age. J Am Acad Orthop Surg 2001; 9:401–411.
7. Lalonde FD, Frick SL, Wenger DR. Surgical correction of residual hip dysplasia
in two pediatric age-groups. J Bone Joint Surg Am 2002; 84-a:1148–1156.
8. Novais EN, Potter GD, Clohisy JC, Millis MB, Kim YJ, Trousdale RT, et al. Obesity
is a major risk factor for the development of complications after peri-acetabular osteotomy. Bone Joint J 2015; 97-b:29–34.
9. Novais EN, Potter GD, Sierra RJ, Kim YJ, Clohisy JC, Schoenecker PL, et al. Surgical treatment of adolescent acetabular dysplasia with a periacetabular osteotomy: does obesity
increase the risk of complications? J Pediatr Orthop 2015; 35:561–564.
10. van der Baan-Slootweg O, Benninga MA, Beelen A, van der Palen J, Tamminga-Smeulders C, Tijssen JG, et al. Inpatient treatment of children and adolescents with severe obesity
in the Netherlands: a randomized clinical trial. JAMA Pediatr 2014; 168:807–814.
12. Barlow SE, Expert C. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity
: summary report. Pediatrics 2007; 120 (Suppl 4):164–192.
14. Saito JM, Chen LE, Hall BL, Kraemer K, Barnhart DC, Byrd C, et al. Risk-adjusted hospital outcomes for children’s surgery. Pediatrics 2013; 132:677–688.
17. McNutt LA, Wu C, Xue X, Hafner JP. Estimating the relative risk in cohort studies and clinical trials of common outcomes. Am J Epidemiol 2003; 157:940–943.
18. Greenland S. Model-based estimation of relative risks and other epidemiologic measures in studies of common outcomes and in case-control studies. Am J Epidemiol 2004; 160:301–305.
19. Basques BA, Lukasiewicz AM, Samuel AM, Webb ML, Bohl DD, Smith BG, Grauer JN. Which pediatric orthopaedic procedures have the greatest risk of adverse outcomes? J Pediatr Orthop 2017; 37:429–434.
20. Basques BA, McLynn RP, Fice MP, Samuel AM, Lukasiewicz AM, Bohl DD, et al. Results of database studies in spine surgery can be influenced by missing data. Clin Orthop Relat Res 2017; 475:2893–2904.
21. Lavoie J. Blood transfusion risks and alternative strategies in pediatric patients. Paediatr Anaesth 2011; 21:14–24.
22. Bolton-Maggs PH. Transfusion and hemovigilance in pediatrics. Pediatr Clin North Am 2013; 60:1527–1540.
23. Vamvakas EC. White-blood-cell-containing allogeneic blood transfusion and postoperative infection or mortality: an updated meta-analysis. Vox Sang 2007; 92:224–232.
Keywords:Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
hip dysplasia; obesity; National Surgical Quality Improvement Program Pediatric; pelvic osteotomy