Increasing Obesity and Comorbidity in Primary THA Patients
In this study in patients undergoing primary THA, we found increasing rates of obesity, medical and psychological comorbidity and a decreasing rate of RA/inflammatory arthritis as the underlying diagnosis over time. To our knowledge, this is the first study to comprehensively examine time-trends in depression, anxiety, medical comorbidities and underlying diagnosis, overall and in different age groups of patients undergoing primary THA. The younger age group is the fastest growing group for the receipt of THA over time among all age groups. Several study findings deserve further discussion.
A key finding from our study was that the prevalence of obesity, including extreme obesity (BMI ≥40), differed significantly by time and seemed to increase over time. By using the height and weight data prospectively captured in the joint registry, we were able to get accurate estimates of obesity and overcame a key limitation of under-diagnosis of obesity based on the use of ICD-9-CM diagnostic code in all the previous studies. Among the patients undergoing primary THA, obesity increased from 33.2% in 1992–95 to 37.1% for in 2002–05. In the U.S. Medicare population, the prevalence of obesity was reported to increase from 2.2% in 1991 to 7.6% in 2008 and in the NIS from 2.2% in 1990 to 5.0% in 2004. Thus, our study is the first to examine the change in BMI over time in a U.S. cohort undergoing THA. The prevalence of obesity of 37% in our 2002–2005 cohort was higher than the 24.5% reported in 2004 for the general U.S. population. Comparing the prevalence of obesity of 5% in the NIS study using the ICD-9 code to the 37% using height and weight data in our study, highlights the extent of under-coding of obesity using ICD-9-CM codes in administrative databases (by 7-fold relative and 32% absolute difference) and the resulting misclassification in earlier studies. Our cohort's demographic and clinical characteristics are similar to those reported from the U.S. NIS, indicating that our sample is representative. This should be comforting to clinicians who might be worried that our institution does not have a representative population, since the Mayo Clinic provides THA to local residents and is also a referral center. To our knowledge, no systematic changes occurred in referral patterns from primary care physicians to orthopedic surgery or policy for reimbursement for primary THA (except slight uniform reduction in compensation across the health care system) in the U.S. during the study period that would explain the time-trends in these characteristics. Another contribution from our study to the literature was our ability to provide overall and age-specific time-trends in BMI and to examine time-trends in extreme obesity.
An important finding from our study was the doubling of the rates of extreme obesity in 50- < 65 and 65- < 80 age groups and of depression in all 4 age groups. To our knowledge this is the first study to examine these important patient characteristics in various age groups, over time. Examination of these characteristics (extreme obesity, depression etc.) by age groups provides a detailed knowledge of their variation over time by age in primary THA patients. This knowledge can also allow policy makers to direct their efforts towards the group that is most at risk due to rapidly increasing prevalence of these clinical features that are associated with poor prognosis and worse outcomes after primary THA.
We also found that the prevalence of RA/inflammatory arthritis reduced by 65–75% in <50 and 65- < 80 age groups over the study period. This finding adds to the growing evidence that fewer patients with RA/inflammatory arthritis are requiring THA, compared to 1–2 decades ago. This may at least partially be due to the availability and more frequent use of effective disease-modifying treatments for RA, including methotrexate and/or biologics, as recommended in the RA treatment guidelines.
The time-trends in the prevalence of extreme obesity, depression and anxiety were mostly independent of age, gender and medical comorbidity, as demonstrated in the multivariable-adjusted models. Anxiety and depression are well-recognized predictors of poorer pain and function outcomes after arthroplasty. This is an important finding, since there were significant time trends of decreasing patient age and increasing medical comorbidity over the same time-period. We noted minimal attenuation of odds ratios comparing unadjusted to multivariable-adjusted models, implying that increasing obesity, depression and anxiety in primary THA patients is not the result of increasing medical comorbidity or decreasing age.
In our study, we found rapidly increasing prevalence of moderate to severe renal disease, cerebrovascular disease and dementia over time. To our knowledge, this has not been reported previously in primary THA patients. This finding indicates that surgeons and patients need to be aware of increasing rate of certain comorbidities over time, which may be associated with an increase in peri-operative complications. This needs further study.
Our study confirms several previous findings including a reduction in mean age and increase in Deyo-Charlson index for patients undergoing primary THA. Findings from the study of the U.S. NIS showed that Deyo-Charlson index increased 0.42 to 0.55 (30% increase) from 1998–2008 in patients undergoing THA, similar to the 36% increase noted in our study. The magnitude of increases is lower compared to a 100% increase of comorbidity from 1 to 2 in Medicare sample from 1991 to 2008. This is not surprising due to the inclusion of all-comers in our and the NIS study vs. only patients 65 years and older (with higher comorbidity) in the Medicare sample. Our study also confirms a doubling of the prevalence of diabetes in patients undergoing primary THA over last 10–15 years, as reported in analyses of Medicare (7.1% in 1991–92 to 15.5% in 2007–08) and the NIS data (6.8% in 1990–94 to 11% in 2000–04).
Our study showed that the change in primary THA patient characteristics with increasing obesity and medical and psychological comorbidity is evident. These time-trends need to be accounted in research studies of THA outcomes and when health care plans and policy makers are making the reimbursement and coverage decisions. Data on several important patient characteristics are limited or absent in Medicare and NIS datasets that can be examined more comprehensively using data from total joint registries, similar to ours.
Our study has several limitations that should be taken into account while interpreting study findings. Our cohort study is subject to residual confounding due to study design, despite the adjustment for several important covariates in our multivariable-adjusted analyses. We used ICD-9 diagnostic codes to define anxiety and depression, and therefore these are likely under-reported. In addition, the increasing incidence of anxiety and depression over time might indicate either a better capture/reporting of these or a true increase in its incidence in this patient population. We are unable to distinguish between these two possibilities.
Study strengths include the use of height/weight data for BMI, a representative sample from a well-established U.S. total joint registry, availability of electronically captured data for more than a decade and robustness of our findings across several sensitivity models (age-adjusted vs. age- and sex-adjusted vs. age, sex- and comorbidity-adjusted models). The similarity of our cohort to other published studies of THA as well as the NIS sample, supports the generalizability of these findings.
Discussion
In this study in patients undergoing primary THA, we found increasing rates of obesity, medical and psychological comorbidity and a decreasing rate of RA/inflammatory arthritis as the underlying diagnosis over time. To our knowledge, this is the first study to comprehensively examine time-trends in depression, anxiety, medical comorbidities and underlying diagnosis, overall and in different age groups of patients undergoing primary THA. The younger age group is the fastest growing group for the receipt of THA over time among all age groups. Several study findings deserve further discussion.
A key finding from our study was that the prevalence of obesity, including extreme obesity (BMI ≥40), differed significantly by time and seemed to increase over time. By using the height and weight data prospectively captured in the joint registry, we were able to get accurate estimates of obesity and overcame a key limitation of under-diagnosis of obesity based on the use of ICD-9-CM diagnostic code in all the previous studies. Among the patients undergoing primary THA, obesity increased from 33.2% in 1992–95 to 37.1% for in 2002–05. In the U.S. Medicare population, the prevalence of obesity was reported to increase from 2.2% in 1991 to 7.6% in 2008 and in the NIS from 2.2% in 1990 to 5.0% in 2004. Thus, our study is the first to examine the change in BMI over time in a U.S. cohort undergoing THA. The prevalence of obesity of 37% in our 2002–2005 cohort was higher than the 24.5% reported in 2004 for the general U.S. population. Comparing the prevalence of obesity of 5% in the NIS study using the ICD-9 code to the 37% using height and weight data in our study, highlights the extent of under-coding of obesity using ICD-9-CM codes in administrative databases (by 7-fold relative and 32% absolute difference) and the resulting misclassification in earlier studies. Our cohort's demographic and clinical characteristics are similar to those reported from the U.S. NIS, indicating that our sample is representative. This should be comforting to clinicians who might be worried that our institution does not have a representative population, since the Mayo Clinic provides THA to local residents and is also a referral center. To our knowledge, no systematic changes occurred in referral patterns from primary care physicians to orthopedic surgery or policy for reimbursement for primary THA (except slight uniform reduction in compensation across the health care system) in the U.S. during the study period that would explain the time-trends in these characteristics. Another contribution from our study to the literature was our ability to provide overall and age-specific time-trends in BMI and to examine time-trends in extreme obesity.
An important finding from our study was the doubling of the rates of extreme obesity in 50- < 65 and 65- < 80 age groups and of depression in all 4 age groups. To our knowledge this is the first study to examine these important patient characteristics in various age groups, over time. Examination of these characteristics (extreme obesity, depression etc.) by age groups provides a detailed knowledge of their variation over time by age in primary THA patients. This knowledge can also allow policy makers to direct their efforts towards the group that is most at risk due to rapidly increasing prevalence of these clinical features that are associated with poor prognosis and worse outcomes after primary THA.
We also found that the prevalence of RA/inflammatory arthritis reduced by 65–75% in <50 and 65- < 80 age groups over the study period. This finding adds to the growing evidence that fewer patients with RA/inflammatory arthritis are requiring THA, compared to 1–2 decades ago. This may at least partially be due to the availability and more frequent use of effective disease-modifying treatments for RA, including methotrexate and/or biologics, as recommended in the RA treatment guidelines.
The time-trends in the prevalence of extreme obesity, depression and anxiety were mostly independent of age, gender and medical comorbidity, as demonstrated in the multivariable-adjusted models. Anxiety and depression are well-recognized predictors of poorer pain and function outcomes after arthroplasty. This is an important finding, since there were significant time trends of decreasing patient age and increasing medical comorbidity over the same time-period. We noted minimal attenuation of odds ratios comparing unadjusted to multivariable-adjusted models, implying that increasing obesity, depression and anxiety in primary THA patients is not the result of increasing medical comorbidity or decreasing age.
In our study, we found rapidly increasing prevalence of moderate to severe renal disease, cerebrovascular disease and dementia over time. To our knowledge, this has not been reported previously in primary THA patients. This finding indicates that surgeons and patients need to be aware of increasing rate of certain comorbidities over time, which may be associated with an increase in peri-operative complications. This needs further study.
Our study confirms several previous findings including a reduction in mean age and increase in Deyo-Charlson index for patients undergoing primary THA. Findings from the study of the U.S. NIS showed that Deyo-Charlson index increased 0.42 to 0.55 (30% increase) from 1998–2008 in patients undergoing THA, similar to the 36% increase noted in our study. The magnitude of increases is lower compared to a 100% increase of comorbidity from 1 to 2 in Medicare sample from 1991 to 2008. This is not surprising due to the inclusion of all-comers in our and the NIS study vs. only patients 65 years and older (with higher comorbidity) in the Medicare sample. Our study also confirms a doubling of the prevalence of diabetes in patients undergoing primary THA over last 10–15 years, as reported in analyses of Medicare (7.1% in 1991–92 to 15.5% in 2007–08) and the NIS data (6.8% in 1990–94 to 11% in 2000–04).
Our study showed that the change in primary THA patient characteristics with increasing obesity and medical and psychological comorbidity is evident. These time-trends need to be accounted in research studies of THA outcomes and when health care plans and policy makers are making the reimbursement and coverage decisions. Data on several important patient characteristics are limited or absent in Medicare and NIS datasets that can be examined more comprehensively using data from total joint registries, similar to ours.
Our study has several limitations that should be taken into account while interpreting study findings. Our cohort study is subject to residual confounding due to study design, despite the adjustment for several important covariates in our multivariable-adjusted analyses. We used ICD-9 diagnostic codes to define anxiety and depression, and therefore these are likely under-reported. In addition, the increasing incidence of anxiety and depression over time might indicate either a better capture/reporting of these or a true increase in its incidence in this patient population. We are unable to distinguish between these two possibilities.
Study strengths include the use of height/weight data for BMI, a representative sample from a well-established U.S. total joint registry, availability of electronically captured data for more than a decade and robustness of our findings across several sensitivity models (age-adjusted vs. age- and sex-adjusted vs. age, sex- and comorbidity-adjusted models). The similarity of our cohort to other published studies of THA as well as the NIS sample, supports the generalizability of these findings.
SHARE
