Impact of age and comorbidity burden on mortality and major complications in older adults undergoing orthopaedic surgery: an analysis using the Japanese diagnosis procedure combination database

The Diagnosis Procedure Combination (DPC) is a case-mix patient classification system launched in 2002 by the Ministry of Health, Labour and Welfare of Japan, and linked with a lump-sum payment system[11‐14]. All 82 university teaching hospitals are obliged to use this system, but adoption by community hospitals is voluntary. The DPC Research Group surveys participating hospitals between July 1 and December 31 each year in collaboration with the Ministry of Health, Labour and Welfare. In 2010, 926 hospitals participated and provided data for 2.9 million patients, approximately 45% of all inpatient admissions to acute care hospitals in Japan. The database includes the International Classification of the Diseases 10th Revision (ICD-10) codes for primary and secondary diagnoses; comorbid conditions that existed at admission; and complications that occurred after admission. To optimize the accuracy of the recorded diagnoses, physicians in charge are obliged to record the diagnoses with reference to medical charts.

The Institutional Review Board at The University of Tokyo approved the study design and waived informed consent because the data is anonymous.

We included all patients aged 50 years or older who had undergone one of the following operations from 2007 to 2010: cervical laminoplasty, lumbar decompression, lumbar arthrodesis, and primary TKA. These procedures were chosen because they are predominantly performed for degenerative conditions commonly seen in the elderly. Although these procedures might be conducted in an urgent manner, this study did not exclude emergency cases. We excluded total hip arthroplasty (THA) from our analysis because of demographic and etiologic differences; the majority of Japanese patients present with osteoarthritis secondary to dysplasia of the hip joint, which is not necessarily the case in other parts of the world.

The variables abstracted from the DPC database were: age, sex, comorbid conditions that existed at admission, surgical site, length of stay, and postoperative adverse outcomes. To evaluate the impact of multiple comorbidities, we used Charlson comorbidity index (CCI) base on Quan’s protocol[15, 16], a weighted index that takes into account the number and the seriousness of comorbid conditions. In calculation of the CCI, patients’ comorbid conditions are classified into following 17 categories: myocardial infarction; congestive heart failure; peripheral vascular disease; cerebrovascular disease; dementia; chronic pulmonary disease; peptic ulcer diease; mild liver disease; diabetes without chronic complication; diabetes with chronic complication; hemiplesia or paraplegia; renal disease; any malignancy, including leukemia and lymphoma; moderate or severe liver disease; metastatic solid tumor; and AIDS/HIV. Each condition is assigned a score of 1, 2, 3, or 6, depending on the risk of dying associated with the condition. Scores are then summed to provide a total score of the CCI. In the present study, coded comorbidities in the DPC database were converted to designated points according to Quan’s protocol.

Primary outcomes included in-hospital death or any of the following complications: surgical site infection (ICD10 code: T793, T814), sepsis (A40, A41), cardiac events (acute coronary events [I21-I24] or heart failure [I50]), respiratory complications (pneumonia [J12-J18], post procedural respiratory disorders [J95] or respiratory failure [J96]), pulmonary embolism (I26), stroke (cerebral infarction or hemorrhage [I60-I64]), acute renal failure [N17].

We used analysis of variance or Kruskal-Wallis tests to compare continuous data; Chi-square tests to compare categorical data; Fisher’s exact test to compare in-hospital mortality and major complication rates between subgroups; logistic regression to analyze concurrent effects of factors on the occurrence of in-hospital deaths and postoperative complications, and a generalized estimating equation to adjust for the clustering of patients within hospitals. In the regressions, in-hospital deaths and postoperative complications were modeled as functions of age, sex, CCI, and surgical procedure[17]. The threshold for significance was < 0.05.

This study is the largest (study population of 107,104) that analyzes the risks associated with current orthopaedic surgery in older adults. The DPC database is a large administrative database similar to National Inpatient Sample in the United States, and the data allows nationwide investigation and comparison of mortality and morbidity between stratified subgroups. In addition, the DPC database allows us to assess perioperative complications by differentiating comorbidities at admission and complications after admission.

In common with other studies using administrative data, a degree of misclassification or underreporting of outcome might have occurred. Although we could not verify data for each patient, we presume the level of miscoding is low because DPC data are coded by physicians and subject to an audit. Data are limited as the DPC database does not record some important clinical data, such as the severity of the disease, levels of the arthrodesis, or type of implant used. Despite these limitations, the results presented here provide important national estimates of inpatient morbidity and mortality after orthopaedic surgery.

The values for mortality for spine surgery (0.11%–0.20%) and primary TKA (0.066%) are similar to those reported in other large database studies[6‐9, 19‐21]. In the present study, comorbidity burden (CCI ≥ 3) had the greatest impact both on in-hospital mortality and occurrence of major complications. In the National Surgical Quality Improvement Program, which examined outcomes for 3,475 patients undergoing spine surgery, identified age, and contaminated or infected wounds as independent predictors of mortality[10]. Pumberger et al. reported the highest odds for perioperative mortality for patients undergoing lumbar arthrodesis were among patients aged 75 years or older (OR, 4.35; reference: 45–46 years) and those with the comorbidities of congestive heart failure, coagulopathy, and liver disease[9]. Similarly, Memtsoudis et al. found that risk factors for mortality following TKA and THA included age (≥ 75 years; OR, 3.70), male gender, ethnicity, emergency admission, and comorbidity[8].

In these previous studies, each comorbid condition was analyzed separately. In the present study, we used the CCI because in elderly patients often present with multiple coexisting conditions. Our data demonstrate that presentation with coexisting conditions has a striking impact on postoperative adverse outcomes.

Risk profile also varies among surgical procedures. After adjusting for age, sex, and comorbidities, the risk for in-hospital death following lumbar arthrodesis or cervical laminoplasty was twice as high as the risk following TKA. There was no significant difference between lumbar decompression and TKA. Few previous studies have examined procedure-related risk differences among orthopaedic procedures. Memtsoudis et al. reported that TKA has a slightly lower risk of mortality than THA[8]. Deyo et al. demonstrated that lumbar arthrodesis, in particular complex fusion, compared with decompression alone, increased risk of 30-day mortality[5]. Although unadjusted factors may influence mortality rates, surgeons should be aware of the different risk profile among surgical procedures.