Perioperative all-cause mortality and cardiovascular events in patients with rheumatoid arthritis: Comparison with unaffected controls and persons with diabetes mellitus

Authors


Abstract

Objective

Rheumatoid arthritis (RA) is associated with an increased cardiovascular (CV) burden similar to that of diabetes mellitus (DM). This risk may warrant preoperative CV assessment as is performed for patients with DM. We aimed to determine whether the risks of perioperative death and CV events among patients with RA differed from those among unaffected controls and patients with DM.

Methods

We used 1998–2002 data from the Nationwide Inpatient Sample (NIS) database of the Healthcare Cost Utilization Project (HCUP) to identify hospitalizations of patients undergoing elective noncardiac surgery. Using established guidelines, surgical procedures were categorized as either low risk, intermediate risk, or high risk of having CV events. Logistic models provided the adjusted odds of study end points in patients with RA, DM, or both relative to patients with neither condition.

Results

Among 7,756,570 patients undergoing a low-risk, intermediate-risk, or high-risk noncardiac procedure, 2.34%, 0.51%, and 2.12%, respectively, had a composite CV event, and death occurred in 1.47%, 0.50%, and 2.59%, respectively. Among those undergoing an intermediate-risk procedure, death was less likely in RA patients than in DM patients (0.30% versus 0.65%; P < 0.001), but the difference in mortality rates among those undergoing low-risk versus high-risk procedures was not significant. Patients with RA were less likely to have a CV event than were patients with DM for procedures of low risk (3.38% versus 5.30%; P < 0.001) and intermediate risk (0.34% versus 1.07%; P < 0.001). In adjusted models, RA was not independently associated with an increased risk of perioperative death or a CV event.

Conclusion

RA was not associated with adverse perioperative CV risk or mortality risk, which suggests that current perioperative clinical care does not need to be changed in this regard.

Persons with rheumatoid arthritis (RA) have a higher risk of death from all causes as well as from cardiovascular (CV)–specific causes as compared to the non-RA population (1, 2). In addition, individuals with RA are at increased risk of experiencing myocardial infarction, congestive heart failure, stroke, and peripheral vascular disease (3–6). Moreover, there is evidence of clinically silent cardiovascular disease in individuals with RA (7). Since the differences in traditional CV risk factors do not entirely explain this elevated risk, the underlying cause of the increased CV risk is likely to be multifactorial. In addition to the traditional CV risk factors, the effects of systemic inflammation, plaque instability, impaired coronary flow reserve, elevated thrombotic markers, and insulin resistance may play a role (7–15). Recent studies have compared the elevated CV risk in RA patients to that in patients with diabetes mellitus (DM) (16, 17). Due to the increase in CV disease burden, there is a recognized need for CV risk assessment and management in RA (18).

Cardiac complications are the major cause of perioperative morbidity and mortality (19, 20). The risk of a perioperative coronary event or cardiac-specific death is nearly two-fold higher in patients with known or suspected coronary or atherosclerotic disease relative to those without (21–23). The presence of increased atherosclerotic burden in patients with RA adds complexity in the perioperative risk assessment. Additionally, physical inactivity and limitation due to RA may make angina an insensitive marker of CAD, as physical exertion may not be at a sufficient level for patients to develop symptoms of angina. Indeed, if RA is a “diabetic equivalent,” then the associated CV risk would need to be incorporated into the preoperative risk stratification.

It is still to be determined whether the increased atherosclerotic burden of RA translates into an increased risk of adverse perioperative events, including CV events and mortality. Additionally, given the recent literature drawing similarities between the CV risk in RA and DM in the nonoperative setting, it is of clinical interest to determine whether these similarities extend to the perioperative setting. Such similarities would imply the need for a more comprehensive approach to preoperative clearance, with a focus on CV risk in patients with RA.

In the present study, our aim was to determine and to compare the risk of perioperative CV events and mortality during elective hospitalizations among patients with RA, DM, both RA and DM, and neither RA nor DM. We hypothesized that patients with RA would have a risk of perioperative CV events and death that would be higher than that of patients without RA or DM and similar to that of patients with DM.

PATIENTS AND METHODS

Study population.

The Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP), which is sponsored by the Agency for Healthcare Research and Quality (AHRQ), is a survey design–based database of discharge information for inpatient care from non-federal (excludes Veterans Hospitals and other federal facilities), nonrehabilitation, acute-care, short-term hospitals (24). The NIS is an annual sample of hospital discharges that provides national estimates of the characteristics of the patients, diagnoses, and hospital-based procedures performed in US acute-care hospitals.

Each year of NIS data used in our analysis included ∼6.8 million hospital discharges, based on a sample from more than 984 hospitals across 22 states. The numbers of hospitals and states included in our analysis increased annually, from 984 hospitals among 22 states in 1998 to 995 hospitals from 35 states in 2002. The data for this study, including the annual NIS datasets, Procedure Class software tool, Clinical Classifications Software (CCS) tool, and NIS Trends Supplemental Files for pooled analysis of multiple years of NIS data, was provided by the HCUP NIS database (24–27).

Inclusion and exclusion criteria.

The characteristics of the subpopulation of interest included the following: hospital discharges between 1998 through 2002 with a noncardiac principal procedure during the hospitalization, patient age 18 years or older, and the hospitalization designated as elective. We included only elective hospitalizations because this group is most relevant for the question of preoperative clearance and because this criterion would ensure the greatest likelihood that the procedure preceded the perioperative CV event.

Covariate definitions.

Medical diagnoses of RA, DM, and other comorbid disease states, including hypertension, congestive heart failure without pulmonary edema, valvular heart disease, coronary artery disease, and chronic kidney disease, were abstracted from among 15 possible NIS discharge diagnosis fields, using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) medical diagnosis codes (see Appendix A). Likewise, the principal procedure during the hospitalization was abstracted using the ICD-9-CM code that was listed by the NIS as the principal procedure. CCS, a software tool developed by the AHRQ, was used to cluster the principal procedures into meaningful procedure categories (see Appendix B) (25).

The CCS-defined principal procedure categories were then further categorized according to cardiac risk level into low, intermediate, or high cardiac risk, based on the clinical practice guidelines published by the American College of Cardiology/American Heart Association Task Force (28). For example, endoscopic and ophthalmologic procedures and procedures involving the breast were categorized as low risk, while orthopedic, prostatic, other intraabdominal procedures were categorized as intermediate risk. High-risk procedures included peripheral vascular surgery or noncardiac surgical procedures within the cardiothoracic cavity. Any secondary procedure was categorized using the Procedure Class software tool, and a tally count variable was created for minor and major therapeutic secondary procedures (27).

Other covariates.

The additional covariates that were included as possible confounders consisted of the following: age, sex, comorbid diseases, including hypertension, heart failure without pulmonary edema, valvular heart disease, coronary artery disease, and chronic kidney disease, and any secondary minor or major therapeutic procedures.

Outcomes.

The study outcomes included all-cause mortality during the hospitalization and a composite CV event during the hospitalization, which was defined as the occurrence of acute myocardial infarction, non–ST-segment elevation myocardial infarction, congestive heart failure with pulmonary edema, or acute cerebrovascular accident.

Statistical analysis.

A 4-level disease status categorical variable was generated to categorize individuals according to the presence or absence of a diagnosis of RA and a diagnosis of DM (neither RA nor DM present, RA present and DM absent, RA absent and DM present, both RA and DM present). Associations between categorical data were determined using the design-adjusted Rao-Scott chi-square test. Logistic regression was used to determine the odds of the composite CV outcome and all-cause mortality among hospitalized individuals with RA alone, DM alone, or both RA and DM, as compared to those without either RA or DM. Bivariate and logistic regression analyses were performed separately for each level of procedure risk, using subpopulation analyses. In order to exclude case-mix, we conducted procedure-specific supplementary analyses to determine the odds of adverse perioperative events among elective hospitalizations with the principal procedure of total knee replacement. Due to the multifactorial and potentially nonatherosclerotic cause of decompensated heart failure, we conducted additional supplementary analyses in which acute heart failure was excluded as a component of the composite CV end point.

The potential cofounders consisted of age as a continuous variable and the following categorical variables: sex, hypertension, coronary artery disease, valvular heart disease, congestive heart failure without pulmonary edema, chronic kidney disease, and minor and major therapeutic secondary procedures. A significance level of 0.05 with a 2-sided test was used for all hypotheses. All statistical analyses were conducted using Stata 11.1 software (StataCorp), accounting for the survey design features of the NIS in order to provide population-based estimates.

RESULTS

Patient characteristics.

From 1998 to 2002, there were an estimated 7,756,570 discharges for an elective hospitalization in which a noncardiac principal procedure was performed in adults ≥18 years of age. The characteristics of the hospitalized patients in whom a low-risk, intermediate-risk, or high-risk principal procedure was performed are displayed in Tables 1, 2, and 3, respectively, categorized by disease status.

Table 1. Characteristics of elective hospitalizations with a low-risk procedure (n = 5,109,450), by RA and DM status*
 Neither RA nor DM (n = 4,495,197)RA without DM (n = 30,740)DM without RA (n = 577,527)Both RA and DM (n = 5,986)
  • *

    n represents the estimated weighted frequency of observations. Values are the percentage (linearized standard error), except for age, which is the mean ± linearized standard error.

  • P < 0.05 versus the group with neither rheumatoid arthritis (RA) nor diabetes mellitus (DM).

  • P < 0.05 versus the group with DM without RA.

Age, years47.3 ± 0.567.3 ± 0.365.4 ± 0.267.6 ± 0.4
% female73.3 (0.7)76.7 (0.6)54.9 (0.4)74.9 (1.4)
% with coronary artery disease11.0 (0.5)18.4 (0.6)34.7 (0.8)27.8 (1.6)
% with hypertension16.9 (0.5)35.2 (0.7)47.4 (0.4)44.7 (1.6)
% with congestive heart failure5.3 (0.2)11.8 (0.5)19.6 (0.4)19.2 (1.3)
% with valvular heart disease0.6 (0.04)1.2 (0.2)1.2 (0.1)1.3 (0.3)
% with chronic kidney disease0.5 (0.02)0.7 (0.1)2.5 (0.1)1.5 (0.4)
Table 2. Characteristics of hospitalized individuals with an intermediate-risk procedure (n = 2,378,824), by RA and DM status*
 Neither RA nor DM (n = 2,378,825)RA without DM (n = 25,787)DM without RA (n = 244,170)Both RA and DM (n = 3,051)
  • *

    n represents the estimated weighted frequency of observations. Values are the percentage (linearized standard error), except for age, which is the mean ± linearized standard error.

  • P < 0.05 versus the group with neither rheumatoid arthritis (RA) nor diabetes mellitus (DM).

  • P < 0.05 versus the group with DM without RA.

Age, years55.9 ± 0.262.7 ± 0.264.6 ± 0.264.2 ± 0.5
% female64.6 (0.4)76.2 (0.6)53.9 (0.4)72.2 (1.8)
% with coronary artery disease6.4 (0.2)8.2 (0.4)19.6 (0.4)15.5 (1.5)
% with hypertension26.9 (0.3)37.5 (0.8)56.7 (0.4)54.6 (1.9)
% with congestive heart failure1.9 (0.05)2.6 (0.3)6.6 (0.2)6.4 (1.1)
% with valvular heart disease0.3 (0.02)0.4 (0.1)0.4 (0.03)0 (0)
% with chronic kidney disease0.3 (0.03)0.2 (0.1)1.5 (0.1)0.6 (0.3)
Table 3. Characteristics of hospitalized individuals with a high-risk procedure (n = 268,295), by RA and DM status*
 Neither RA nor DM (n = 192,961)RA without DM (n = 1,946)DM without RA (n = 72,827)Both RA and DM (n = 561)
  • *

    n represents the estimated weighted frequency of observations. Values are the percentage (linearized standard error), except for age, which is the mean ± linearized standard error.

  • P < 0.05 versus the group with neither rheumatoid arthritis (RA) nor diabetes mellitus (DM).

  • P < 0.05 versus the group with DM without RA.

Age, years66.0 ± 0.669.1 ± 0.766.6 ± 0.265.9 ± 1.1
% female41.5 (0.3)64.1 (2.8)43.1 (0.6)56.8 (5.6)
% with coronary artery disease31.1 (0.8)31.8 (2.4)38.1 (0.9)28.9 (4.6)
% with hypertension42.7 (0.6)48.5 (0.2)47.0 (0.8)47.6 (5.3)
% with congestive heart failure12.5 (0.3)11.1 (1.6)16.2 (0.4)12.7 (3.2)
% with valvular heart disease5.2 (0.3)6.4 (1.2)2.2 (0.2)4.3 (1.9)
% with chronic kidney disease2.6 (0.2)1.5 (0.6)6.1 (0.5)2.6 (1.5)

RA patients were significantly older and were more likely to be female as compared to patients with DM and patients without either DM or RA, regardless of the procedure risk category. In general, the comorbid conditions of coronary artery disease, hypertension, congestive heart failure, and chronic kidney disease were significantly more common in patients with DM than those with RA alone or those with neither RA nor DM. One exception to this general pattern was that the proportion of those with DM and those with RA who had hypertension did not differ significantly among those with high-risk procedures. Also, the difference in the proportion of valvular heart disease in patients with DM and those with RA was not statistically significant among the low-risk and intermediate-risk categories.

Death during hospitalization.

The frequency of perioperative all-cause mortality by procedure risk level and disease status is displayed in Table 4. Among patients with DM alone, those who underwent a principal procedure considered low risk (unadjusted odds ratio [OR] 1.54 [95% confidence interval (95% CI) 1.23–1.93]) or intermediate risk (unadjusted OR 1.34 [95% CI 1.19–1.51]) were more likely to die during that hospitalization as compared to those without RA or DM; this risk was attenuated after sequentially adjusting for age and sex, for potential confounders, and for secondary procedures (models B, C, and D, respectively) as displayed in Table 5. Persons with RA alone with a procedure of any risk level were not at a greater risk of all-cause mortality as compared to those without RA or DM.

Table 4. All-cause mortality and composite cardiovascular events occurring overall as well as by disease status and procedure risk level*
 Procedure risk
LowIntermediateHigh
  • *

    Values are the frequency (%) of events.

  • P < 0.05 versus the group with neither rheumatoid arthritis (RA) nor diabetes mellitus (DM).

All-cause mortality   
 Overall75,162 (1.47)11,828 (0.50)6,947 (2.59)
 Disease status   
  Neither RA nor DM62,254 (1.39)10,153 (0.48)5,514 (2.86)
  RA without DM544 (1.77)78 (0.30)40 (2.03)
  DM without RA12,229 (2.12)1,578 (0.65)1,393 (1.91)
  Both RA and DM135 (2.24)19 (0.62)0 (0)
Composite cardiovascular event   
 Overall119,727 (2.34)12,017 (0.51)5,675 (2.12)
 Disease status   
  Neither RA nor DM87,807 (1.95)9,283 (0.44)4,145 (2.15)
  RA without DM1,038 (3.38)88 (0.34)31 (1.58)
  DM without RA30,636 (5.30)2,601 (1.07)1,490 (2.05)
  Both RA and DM246 (4.1)45 (1.47)9 (1.68)
Table 5. Risk of all-cause mortality by disease status and principal procedure risk level*
 Disease status
Neither RA nor DMRA without DMDM without RABoth RA and DM
  • *

    Model A represents the unadjusted model. Model B represents the model adjusted for age and sex. Model C includes model B covariates plus comorbid diseases (coronary artery disease, hypertension, congestive heart failure, valvular heart disease, and chronic kidney disease). Model D includes model C covariates plus minor and major secondary therapeutic procedures. Values are the odds ratio (95% confidence interval). RA = rheumatoid arthritis; DM = diabetes mellitus.

Low-risk procedures    
 Model A1.001.28 (0.94–1.76)1.54 (1.23–1.93)1.64 (1.04–2.57)
 Model B1.000.70 (0.51–0.96)0.83 (0.66–1.04)0.88 (0.56–1.40)
 Model C1.000.69 (0.50–1.95)0.85 (0.73–0.99)0.89 (0.57–1.37)
 Model D1.000.69 (0.50–0.95)0.84 (0.72–0.99)0.88 (0.57–1.36)
Intermediate-risk procedures    
 Model A1.000.62 (0.39–1.01)1.34 (1.19–1.51)1.29 (0.50–3.37)
 Model B1.000.55 (0.34–0.88)0.96 (0.85–1.09)1.08 (0.41–2.82)
 Model C1.000.55 (0.34–0.89)0.93 (0.83–1.06)1.08 (0.40–2.90)
 Model D1.000.60 (0.37–0.97)0.98 (0.86–1.11)1.20 (0.45–3.19)
High-risk procedures    
 Model A1.000.70 (0.35–1.39)0.66 (0.57–0.76)
 Model B1.000.63 (0.32–1.25)0.67 (0.58–0.78)
 Model C1.000.67 (0.33–1.32)0.67 (0.58–0.76)
 Model D1.000.70 (0.35–1.37)0.69 (0.60–0.79)

Composite cardiovascular outcome.

The frequency of perioperative composite CV events by procedure risk level and disease status is displayed in Table 4. In addition to the group comparisons shown in Table 4, we performed further group analyses comparing persons with RA and DM to persons without RA but with DM. These comparisons revealed that persons without RA but with DM who underwent low-risk procedures had greater composite CV events, whereas the differences among those who underwent intermediate-risk or high-risk procedures were not significant.

The risk of composite CV events occurring among those with a low-, intermediate-, or high-risk principal procedure is displayed in Table 6. In an unadjusted logistic model (model A), patients with RA alone who underwent a low-risk principal procedure had nearly 2-fold greater odds of experiencing a composite CV event as did those without RA or DM. Persons with DM alone had more than 2.5-fold greater odds of experiencing a composite CV event. The increased odds of perioperative CV events observed in those with RA alone was significantly attenuated after adjusting for age and sex (model B), while the increased composite CVD risk in persons with DM alone remained statistically significant in a fully adjusted model (model D). A similar pattern of association was observed among persons with an intermediate-risk procedure. Among persons with a high-risk principal procedure, neither those with RA alone nor those with DM alone had a significantly greater risk of a composite CV event relative to those without RA or DM.

Table 6. Risk of composite cardiovascular event by disease status and principal procedure risk level*
 Disease status
Neither RA nor DMRA without DMDM without RABoth RA and DM
  • *

    Model A represents the unadjusted model. Model B represents the model adjusted for age and sex. Model C includes model B covariates plus comorbid diseases (coronary artery disease, hypertension, congestive heart failure, valvular heart disease, and chronic kidney disease). Model D includes model C covariates plus minor and major secondary therapeutic procedures. Values are the odds ratio (95% confidence interval). RA = rheumatoid arthritis; DM = diabetes mellitus.

Low-risk procedures    
 Model A1.001.75 (1.47–2.09)2.81 (2.63–3.01)2.15 (1.58–2.91)
 Model B1.001.00 (0.86–1.18)1.51 (1.44–1.58)1.18 (0.87–1.61)
 Model C1.001.03 (0.88–1.20)1.17 (1.12–1.22)0.99 (0.72–1.35)
 Model D1.001.03 (0.88–1.20)1.16 (1.11–1.22)0.98 (0.72–1.35)
Intermediate-risk procedures    
 Model A1.000.78 (0.46–1.31)2.43 (2.21–2.68)3.30 (1.67–6.54)
 Model B1.000.64 (0.38–1.07)1.70 (1.54–1.87)2.60 (1.30–5.19)
 Model C1.000.63 (0.37–1.06)1.32 (1.19–1.46)2.13 (1.07–4.21)
 Model D1.000.64 (0.38–1.08)1.33 (1.20–1.48)2.17 (1.10–4.29)
High-risk procedures    
 Model A1.000.73 (0.33–1.63)0.95 (0.83–1.09)0.78 (0.19–3.10)
 Model B1.000.69 (0.31–1.54)0.94 (0.82–1.08)0.79 (0.20–3.14)
 Model C1.000.74 (0.33–1.66)0.89 (0.77–1.03)0.80 (0.19–3.24)
 Model D1.000.76 (0.34–1.72)0.92 (0.80–1.05)0.84 (0.20–3.47)

The results of our supplementary analyses did not alter the conclusion based on our primary analysis. The results of procedure-specific analyses to determine the odds of adverse perioperative events among patients undergoing elective hospitalizations with a principal procedure of total knee replacement were consistent with the results of the procedure group–based analyses. Relative to persons without RA or DM, persons with RA alone (adjusted OR 0.38 [95% CI 0.05–2.92]) or persons with DM alone (adjusted OR 1.10 [95% CI 0.60–2.02]) were not at increased risk of all-cause mortality, while persons with both RA and DM (adjusted OR 6.49 [95% CI 1.45–29.04]) were at significantly increased risk of all-cause mortality. Relative to persons with neither DM nor RA, those who had RA alone were not at greater risk of a composite CV event (adjusted OR 0.47 [95% CI 0.14–1.53]). In contrast, persons with DM were at increased risk of a composite CV event, which was observed among those with DM but without RA (adjusted OR 1.51 [95% CI 1.06–2.14]) as well as those with RA (adjusted OR 3.50 [95% CI 1.10–11.19]).

DISCUSSION

In this large population-based study of hospitalized individuals, we sought to determine the association between the absence or presence of either or both diagnoses of RA or DM and perioperative CV events as well as all-cause mortality. In our analyses, we found that hospitalized RA patients who underwent a noncardiac principal procedure were not at an increased risk of in-hospital short-term perioperative adverse CV events or death as compared to hospitalized patients without RA or DM. These findings were consistent across all 3 levels of procedural risk and after controlling for comorbid conditions and additional secondary hospital-based procedures. To our knowledge, this is the first study to investigate the association between RA and perioperative risk of death and CV events.

Recent studies comparing the risk of CV disease in RA with DM patients have found an elevated risk of CV events in RA that was comparable in magnitude to the risk in DM (16, 17). In the perioperative setting, we compared the prevalence of CV comorbid conditions and the risk of perioperative CV events among hospitalized persons with either RA, DM, neither RA nor DM, or both RA and DM. In comparison to persons without RA or DM, persons with RA did indeed have a higher prevalence of CV comorbid diseases, including coronary artery disease, hypertension, and congestive heart failure. With regard to the risk of adverse perioperative events, our results were consistent with the published literature, in that persons with DM remained at an increased risk of adverse perioperative events (29, 30). However, despite the higher prevalence of CV comorbid diseases, our study did not show that persons with RA alone had an elevated risk of CV events. Given the high prevalence of CV comorbid conditions in persons with RA in our study population, we addressed the possibility of model overadjustment by including a multivariable model that adjusted for only age and sex. However, the results of this intermediate multivariable model did not reveal that overadjustment accounted for the lack of association between RA and adverse perioperative events.

The results of our supplementary analyses were in large part consistent with the results of our primary analyses. In order to account for the possibility of procedure case-mix within each procedure risk category, we conducted a separate procedure-specific analysis to determine the odds of adverse perioperative events among elective hospitalizations with a principal procedure of total knee replacement. While the result of this supplementary analysis was consistent among persons with RA alone, it did suggest that the subgroup of persons with both RA and DM may be at heightened risk of experiencing adverse perioperative events. Due to the multifactorial and possibly nonatherosclerotic cause of decompensated heart failure, we conducted additional analyses excluding acute heart failure as a component of the composite CV end point. The results of this analysis did alter our conclusion.

Given the large body of evidence suggesting an increased burden of CV disease in patients with RA, we consider some limitations and potential explanations for our somewhat surprising findings. First, our study was limited due to factors related to its observational and cross-sectional design.

Second, some possible explanations for our results include a differential prehospitalization preoperative risk assessment and modification. Specifically, patients with active or aggressive RA would probably be less likely to undergo an elective noncardiac surgical procedure. RA patients may not have been referred for an elective surgical procedure because of the recognition of the overall increased risk of CV events, which would lead to a selection process that excluded RA patients from elective surgical procedures. Given that much of the epidemiologic research highlighting the association between RA and CV disease has been carried out during the last decade and therefore postdates the time of this dataset, which was 1998–2002, this possibility seems unlikely. Likewise, patients with DM who at greater perioperative risk may have been identified prior to an elective surgery because of preoperative scrutiny, thus underestimating the risk in these patients relative to an unselected group of patients with DM.

Third, a number of RA-specific factors that were not available in the NIS dataset, such as the disease duration, magnitude of systemic inflammation, use of disease-modifying drugs, as well as functional status, may have had an effect on the CV disease profile (4, 31–33). Fourth, one must consider the possibility of coding errors or misclassification when using ICD-9 codes for abstraction of diseases and procedures (34–36). For example, the misclassification of a noninflammatory arthritis as RA may have led to an attenuation of the association between the disease and perioperative CV events.

Fifth, while a mix of procedures within each procedure risk group, especially if these differed by diagnosis, may have introduced residual confounding, the results of our supplementary analysis make this possibility less likely. Last, there was a lack of an anticipated graded increase in the rate of adverse perioperative events across low-, intermediate-, and high-risk procedures that may have been due to the above-mentioned possibilities of differential preoperative screening and limitations inherent to the use of ICD-9 codes as well as other limitations, such as the lack of additional information on procedure-related factors, including indication and duration.

The strengths of our study include the large population-based sample of patients, the comparison with both unaffected controls and patients with DM, and examination of risks within surgical risk strata. Our study serves as the initial step in attempting to quantify the significance of the increased atherosclerotic burden in patients with RA during the perioperative period, serving well those who encounter this task routinely in clinical practice. Several risk stratification indices have been developed to aid clinicians in stratifying preoperative risk, with their components including various recognized CV disease states, such as coronary heart disease, congestive heart failure, and valvular heart disease (28). These indices include characteristics such as age and the presence or absence of DM, chronic kidney disease, and hypertension. Accumulating evidence of the CV disease burden in RA prompted our investigation of the risk associated with RA in the perioperative setting (18).

Our findings did not reveal a significant association between RA and either death or perioperative CV events. The results of our cross-sectional observational study presented here do not support the need for a change in the preoperative screening for CV disease as currently practiced in persons with RA who are being evaluated before elective noncardiac procedures. Our results reflect short-term in-hospital adverse outcomes, however, and do not exclude the possibility of an increased risk of adverse postoperative events after a longer followup time after surgery. Therefore, prospective studies are warranted to further elucidate the relationship between RA and perioperative CV risk.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Yazdanyar had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Yazdanyar, Wasko, Kraemer, Ward.

Acquisition of data. Yazdanyar.

Analysis and interpretation of data. Yazdanyar, Wasko, Kraemer, Ward.

APPENDIX A

Table  . ICD-9-CM CODES FOR STUDY OUTCOME AND COMORBID DISEASES
DiagnosisICD-9-CM code
Outcome diagnosis 
 Composite cardiovascular 
  Acute heart failure428.21, 428.23, 428.31, 428.33, 428.41, and 428.43
  Myocardial infarction410.00, 410.01, 410.02, 410.10, 410.11, 410.12, 410.13, 410.20, 410.21, 410.22, 410.23, 410.30, 410.31, 410.32, 410.40, 410.41, 410.42, 410.50, 410.51, 410.52, 410.60, 410.61, 410.62, 410.70, 410.71, 410.72, 410.80, 410.81, 410.82, 410.90, 410.91, and 410.92
  Acute stroke and transient ischemic attack435.8, 435.9, and 436
 Comorbid diseases 
  Coronary artery disease414.00, 414.01, 414.02, 414.03, 414.04, 414.06, 414.07, 414.20, 414.80, and 414.90
  Diabetes mellitus250.00, 250.01, 250.02, 250.03, 250.10, 250.11, 250.12, 250.13, 250.20, 250.21, 250.22, 250.23, 250.30, 250.31, 250.32, 250.33, 250.40, 250.41, 250.42, 250.43, 250.50, 250.51, 250.52, 250.53, 250.60, 250.61, 250.62, 250.63, 250.70, 250.71, 250.72, 250.73, 250.80, 250.81, 250.82, 250.83, 250.90, 250.91, 250.92, and 250.93
  Valvular heart disease394.00, 394.10, 394.20, 394.90, 395.00, 395.10, 395.20, 395.90, 396.10, 396.20, 396.30, 396.80, 396.90, 397.00, and 397.10
  Chronic heart failure428.00, 428.10, 428.20, 428.22, 428.30, 428.32, 428.40, 428.20, and 428.90
  Hypertension410.10, 401.90, 405.11, 405.19, 405.91, and 405.99
  Chronic kidney disease585.0, 585.1, 585.2, 585.4, 585.5, 585.6, and 585.9

APPENDIX B

Table  . SINGLE-LEVEL CCS PROCEDURE CATEGORY LABELS AND ASSOCIATED PROCEDURE RISK LEVEL
Procedure riskCCS label
  • *

    CCS labels 44–50 were excluded because representing cardiac procedures (see ref.25).

LowAll other CCS labels excluding 44–50*
Intermediate1, 2, 3, 10, 36, 42, 51, 57, 59, 66, 67, 71, 72, 73, 74, 75, 78, 79, 80, 84, 89, 94, 96, 99, 103, 104, 106, 112, 113, 114, 119, 120, 124, 125, 132, 152, 153, 154, 158, 161, and 162
High43, 52, 55, 56, 60, 61, 105, 157, and 176

Ancillary