Annals of Surgical Oncology

, Volume 18, Issue 10, pp 2714–2721

Adrenalectomy in Older Americans has Increased Morbidity and Mortality: An Analysis of 6,416 Patients

Authors

  • Hadiza S. Kazaure
    • Division of Endocrine Surgery, Department of SurgeryYale University School of Medicine
    • Division of Endocrine Surgery, Department of SurgeryYale University School of Medicine
  • Julie A. Sosa
    • Division of Endocrine Surgery, Department of SurgeryYale University School of Medicine
Healthcare Policy and Outcomes

DOI: 10.1245/s10434-011-1757-5

Cite this article as:
Kazaure, H.S., Roman, S.A. & Sosa, J.A. Ann Surg Oncol (2011) 18: 2714. doi:10.1245/s10434-011-1757-5
  • 147 Views

Abstract

Background

The incidence of adrenal tumors increases with age. We examined the impact of older age (>60 years) on clinical and economic outcomes after adrenalectomy.

Methods

Adult patients who underwent adrenalectomy in the Healthcare Cost and Utilization Project Nationwide Inpatient Sample (HCUP-NIS) from 2003–2008 were categorized into age groups: ≤60 years, 61–70 years, and >70 years. Outcomes were compared using χ2 and ANOVA; multivariate regression was used to assess the independent effect of older age on adrenalectomy outcomes.

Results

There were 6,416 patients: 21.9% were 61–70 years, and 12.9% were >70 years. Compared with patients ≤60 years, patients 61–70 and >70 years had more complications (14.1% vs. 19.9 and 22.6%; p < 0.001) and mortality (0.4% vs. 1.3 and 2.3%; p < 0.001), longer mean length of stay (LOS) (3.3 vs. 4.0 and 4.9 days; p < 0.001), and higher mean costs ($12,307 vs. $13,226 and $14,649; p < 0.001). After adjustment, older age remained independently associated with sustaining one or more complications after adrenalectomy (odds ratio (OR) 1.4, 95% confidence interval (CI) 1.1–1.7, for patients 61–70 years; OR 1.7, 95% CI 1.3–2.2 for patients >70 years) and longer adjusted LOS (1-day difference, p < 0.01). Age >70 years was independently associated with increased mortality after adrenalectomy (OR 2.8; 95% CI 1.4–5.6). Complications, LOS, and costs were reduced if patients underwent surgery by high-volume compared with low-volume surgeons.

Conclusions

Older age seems to be independently associated with adverse short-term clinical and economic outcomes after adrenalectomy. Enhanced access to high-volume surgeons is a potentially modifiable factor of particular importance in these patients.

According to the U.S. Census Bureau, the number of Americans aged 65 years and older is expected to increase by 120% between the years 2010 and 2050. Americans aged 65 years and older constituted only 13% of the population in 2000, but they are projected to comprise 20% of the U.S. population by 2050.1 It is well known that older individuals require more medical services than their younger counterparts and represent a large and growing part of surgical caseloads.24 Older patients are more likely to sustain surgical complications, which can trigger a cascade of events resulting in loss of independence, high health care costs, and mortality.3 The American Medical Association concluded in a landmark article that the growing elderly population presents new challenges to the health care system.5

Despite the rarity of primary adrenal cancers, adrenal masses are among the most prevalent human tumors.6,7 Adrenal masses increase in frequency with advancing age; autopsy studies indicate that the prevalence of clinically inapparent adrenal masses (incidentalomas) is less than 1% for patients younger than 30 years, and up to 7% in patients aged 70 years or older.811 Improvements in imaging technology have boosted the likelihood of discovering adrenal incidentalomas.911 Their optimal management is an important aspect of patient care.6,11,12 Adrenalectomy is the definitive therapy for most significant adrenal masses.13

Data on age-stratified outcomes of patients who underwent adrenalectomy are limited. In a study examining trends in adrenalectomy outcomes between 1998 and 2006, Murphy et al. suggested that increased frequency of adrenalectomy in older, sicker patients may explain the time trend toward higher complication rates identified in their study.13 This study was designed to analyze the impact of older age (>60 years) on clinical and economic outcomes of adrenalectomy patients at a national level.

Methods

Data Source

The data source was the Nationwide Inpatient Sample (NIS), 2003–2008. Maintained by the Healthcare Cost and Utilization Project (HCUP) of the Agency for Healthcare Research and Quality, the NIS is the largest all-payer inpatient care database in the United States containing information regarding up to eight million hospital stays per year. NIS (2003–2008) consists of a 20% sample of hospital discharges involving approximately 1,000 nonfederal hospitals from more than 35 states. It uses International Classification of Diseases, 9th edition, Clinical Modification (ICD-9-CM) codes, and up to 15 diagnoses and 15 procedures associated with the index hospitalization are included in the database.14

Adults patients (≥18 years) who underwent adrenalectomy (ICD-9-CM codes: 07.2, 07.21, 07.22, 07.29, 07.3) as their primary surgical procedure were abstracted from the NIS database. Patients who underwent concurrent urologic procedures (ICD-9-CM codes: 55.4, 55.5, 55.51, 55.52, 55.54, 55.39, 55.34, 55.53, 55.69, 55.91) were excluded. Patients were categorized into three age groups: ≤60, 61–70, and >70 years. Demographic variables included gender, race, and median household income, represented in quartiles (low, medium-low, medium, and high). Clinical variables were patient comorbidity, functional adrenal disease (hypercortisolism [ICD-9-CM: 255.0], hyperaldosteronism [ICD-9-CM:2551], “medulloadrenal hyperfunction” [ICD-9-CM: 2556]); diagnosis of adrenal malignancy (primary adrenal [ICD-9-CM:194.0] and secondary adrenal malignancies [ICD-9-CM:198.7]), and type of procedure (partial [ICD-9-CM: 07.2,07.21,07.29], unilateral [ICD-9-CM:07.22] or bilateral adrenalectomy [ICD-9-CM:07.3]). Patient comorbidity was assessed by using the enhanced Charlson Comorbidity Index.15 Charlson Comorbidity Index scores 0–1 were categorized as low, 2–3 medium-low, 4–5 moderate, and ≥6 as high. Because there is no unique ICD-9-CM code for laparoscopic adrenalectomy, patients who underwent abdominal laparoscopy (ICD-9-CM:54.21) and adrenalectomy on the same day were identified as patients who underwent laparoscopic adrenalectomy.

Hospital-provider variables were hospital region (northeast, midwest, south, west), location (urban vs. rural), teaching status (teaching vs. non-teaching), and adrenalectomy volume. Surgeon adrenalectomy volume also was analyzed. Hospital and surgeon volume were modeled as dichotomous variables (low- vs. high-volume) based on the number of adrenalectomies performed per year. High-volume hospitals were those above the 75th percentile (>6 adrenalectomies per year) of annual adrenalectomy volume. High-volume surgeons were defined as those who performed four or more adrenalectomies per year, as described by Park et al.16

Outcomes

Clinical outcomes of interest were in-hospital complications, in-hospital mortality, and patient disposition at time of discharge. Economic outcomes were hospital length of stay (LOS) and total inpatient hospital costs. Complications were treated as a binomial outcome (complication vs. no complication). Complications also were categorized into groups—bleeding, pulmonary, cardiovascular, renal, endocrine (corticoadrenal insufficiency), and wound (infectious, wound/technical-related)—as defined by Stavrakis et al.17 Total inpatient costs were calculated using HCUP-NIS adjusted, hospital-specific cost-to-charge ratios. Costs were adjusted for inflation, converting all costs to year 2010 dollar values, using rates from the U.S. Bureau of Labor Statistics.18 Adjusted costs and LOS were then log-transformed to achieve a more normal distribution. Subgroup analyses were performed to examine the effect of surgeon volume on adrenalectomy outcomes by age group.

Statistical Analysis

Bivariate analyses comparing independent variables and outcomes of interest were performed using two-tailed χ2 analysis for categorical variables and two-sided Analysis of Variance (ANOVA) for continuous variables. All p values <0.05 were considered significant. Multivariate logistic and linear regression were used to assess whether older age was independently associated with adverse outcomes after adrenalectomy. A p value <0.2 on bivariate analyses was used to identify preoperative variables entered into our multivariate regression models; a p value <0.05 was the significant criterion used to identify independent risk factors in these models. Odds ratios with 95% confidence intervals (CI) were calculated for final multivariate logistic regression models. Beta coefficients obtained from the final linear regression model were used to estimate the number of adjusted hospital days attributable to older age.

Data analyses and management were performed using SPSS® for Windows version 17.0 software program (SPSS Inc., Chicago, IL). In compliance with confidentiality standards set forth by HCUP, data with less than 11 patients per category are not reported. HCUP-NIS is a public database with no personal identifying information; this study was deemed exempt at our institution from Institutional Review Board approval.

Results

Baseline Characteristics of Adrenalectomy Patients

There were 6,416 patients in the study. Older patients (>60 years) comprised nearly 35% of the study sample; approximately 22 and 13% of the patients were aged 61–70 and > 70 years, respectively (Table 1). The mean age for the ≤60, 61–70, and >70 age groups was 46, 65, and 76 years, respectively. Adrenalectomy patients were more likely to be female and white (p < 0.001 for both) in all age groups.
Table 1

Demographic, clinical, and provider characteristics of patients undergoing adrenalectomy by age group, HCUP-NIS (2003–2008), N = 6,416

Characteristica

Overall

Age (years)

p

≤60

61–70

>70

(n = 4,185) (65.2%)

(n = 1,405) (21.9%)

(n = 826) (12.9%)

Demographic

 Age (years), mean (SEM)

54.4 (0.2)

46.4 (0.2)

65.2 (0.1)

76.2 (0.1)

 

 Gender, female

58.4

60.3

55.2

55.2

<0.001

 Race (n = 4,954)

    

<0.001

  White

76.1

72.9

80.8

84.2

 

  Black

11.1

12.1

10.1

7.6

 

  Hispanic

6.8

7.8

5.5

4.5

 

  Other/unknown

6

7.2

3.6

3.7

 

 Median household income

    

0.15

  Low

20.7

20.3

21.5

21.2

 

  Medium-low

25

24.2

26.7

26.5

 

  Medium

25.3

26

22.9

25.8

 

  High

26.8

27.1

26.8

25.1

 

Clinical

 Charlson Comorbidity Index

    

<0.001

  Low

49.6

59

33.6

28.7

 

  Medium-low

30.2

26.8

37.3

35.7

 

  Moderate

6.1

4

9

11.6

 

  High

14.1

10.2

20.1

24

 

 Functional adrenal disease

8.7

10.7

6

2.9

<0.001

 Malignant adrenal disease

    

<0.001

  Primary

6.7

6.4

6.8

7.7

 

  Secondary

9.5

6.4

14.5

16.6

 

  Laparoscopy

15.4

16.3

14.3

12.7

0.014

  Principal procedure, unilateral

87.7

87.1

88.3

90

0.016

Provider

 Geographic region

    

0.005

  Northeast

21.1

21.4

19.6

21.9

 

  Midwest

18.6

17.9

18.5

22.2

 

  South

38

37.4

40.7

36.8

 

  West

22.3

23.3

21.2

19.1

 

 Urban location

96.1

96.2

96.1

95.5

0.674

 Teaching hospital

72.5

74.6

69.8

66.8

<0.001

 Hospital volume, high

64.9

66

64.3

60.5

0.009

 Surgeon volume, high (n = 3,573)

39.7

41.5

37.2

34.8

0.007

SEM standard error of mean

aAll characteristics except age are expressed as percentages. Percentages have been rounded and may not add up to 100

Older patients were more likely to have higher Charlson comorbidity scores (p < 0.001) and malignant adrenal disease (p < 0.001), but they had lower rates of functional adrenal disease (p < 0.001; Table 1). Approximately 7.5% of the study sample had a diagnosis of hypercortisolism; this diagnosis was less common in older patients (2.2% for patients >70 years, 5.3% for patients aged 61–70 vs. 9.2% for patients ≤60 years; p < 0.001).

Compared with patients ≤60 years, older patients were more likely to have adrenalectomies in nonteaching (p < 0.001) and low-volume hospitals (p = 0.009). Older patients also were more likely to have adrenalectomies by low-volume surgeons (62.8% of patients aged 61–70 years and 65.2% of patients >70 years vs. 58.5% of patients ≤60 years; p = 0.007).

Unadjusted Outcomes

The overall complication rate was 16.5% (Table 2). Compared with a 14.1% complication rate for patients ≤60 years, there was a 41 and 60% increase in complication rates for patients aged 61–70 and >70 years, respectively (p < 0.001).The most common complications in older patients were bleeding, pulmonary, renal, and adrenal insufficiency-related. Only 14.6% of patients who experienced adrenal insufficiency had bilateral adrenalectomy. Among patients who experienced adrenal insufficiency, older patients were less likely to have a diagnosis of hypercortisolism (p = 0.022) but more likely to have malignant adrenal disease (p = 0.047).
Table 2

Unadjusted outcomes following adrenalectomy by age group, HCUP-NIS (2003–2008), N = 6,416

Outcome

Overall

Age (year)

p

≤60

61–70

>70

(n = 4,185) (65.2%)

(n = 1,405) (21.9%)

(n = 826) (12.9%)

Clinical (%)

 Complication, ≥1

16.5

14.1

19.9

22.6

<0.001

 Type of complication

    

<0.001

  Bleeding

6.1

5.3

7.6

7.7

 

  Wound

1.7

1.5

1.9

2.3

 

  Corticoadrenal insufficiency

1.9

1.7

2.0

3

 

  Cardiovascular Renal

1.3

1.2

1.0

2.1

 

  Renal

1.7

1.2

2.3

2.8

 

  Pulmonary

3.8

3.2

5.1

4.7

 

 Mortality

0.8

0.4

1.3

2.3

<0.001

Economic

 Mean length of stay (days)

3.7

3.3

4.0

4.9

<0.001

 Mean cost ($10,000s)

12.8

12.3

13.2

14.7

<0.001

The overall in-hospital mortality rate was 0.8%. Mortality increased with age (p < 0.001), but there was no significant difference in mortality rates with respect to occurrence of adrenal insufficiency (p = 0.382). Among patients who were discharged, older patients were more likely to be discharged to an intermediate care facility (3.7% of patients aged 61–70 and 12.1% of patients >70 years, vs. 1.7% of patients ≤60 years; p < 0.001).

Compared with a mean LOS of 3.3 days for patients ≤60 years, patients aged 61–70 and >70 years experienced a 21 and 48% increase in LOS, respectively. Older age also was associated with higher cost of hospitalization (p < 0.001).

Adjusted Outcomes

After adjusting for demographic, clinical, hospital, and surgeon characteristics, older age remained an independent predictor of sustaining any complication (odds ratio (OR) 1.4, 95% confidence interval (CI) 1.1–1.7 for patients 61–70 years; OR 1.7, 95% CI 1.3–2.2 for patients >70 years; Fig. 1) and longer LOS (1-day difference, p < 0.01) after adrenalectomy. Age >70 years was independently associated with increased risk of mortality (OR 2.8; 95% CI 1.4–5.6) after adrenalectomy. Other predictors of mortality were a moderate (OR 4.1; 95% CI 1.4–11.5) and high (OR 6.1; 95% CI 3.0–12.6) Charlson Comorbidity Index.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-1757-5/MediaObjects/10434_2011_1757_Fig1_HTML.gif
Fig. 1

Multivariate analysis: Independent predictors of any complication after adrenalectomy, HCUP-NIS (2003–2008). Multivariate logistic regression model adjusted for gender, adrenal malignancy, functional adrenal disease, surgical approach, hospital teaching status, and hospital volume. CCI Charlson Comorbidity Index. Referents: age ≤60 years; CCI: low; procedure type: unilateral adrenalectomy; surgical approach: laparoscopy; surgeon volume: high

Subanalysis by Surgeon Volume

We performed a surgeon volume-stratified subanalysis of adrenalectomy outcomes by age group (Figs. 2, 3, 4). Mortality was excluded from the subanalysis because of the small number of occurrences. Patients with surgeon identifiers did not differ from patients without surgeon identifiers with regard to risk factors that were independently associated with morbidity, mortality, and LOS.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-1757-5/MediaObjects/10434_2011_1757_Fig2_HTML.gif
Fig. 2

Adrenalectomy HCUP-NIS (2003–2008). Complication rates of age groups, by surgeon volume

https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-1757-5/MediaObjects/10434_2011_1757_Fig3_HTML.gif
Fig. 3

Adrenalectomy HCUP-NIS (2003–2008). Mean hospital length of stay of age groups, by surgeon volume

https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-1757-5/MediaObjects/10434_2011_1757_Fig4_HTML.gif
Fig. 4

Adrenalectomy HCUP-NIS (2003–2008). Mean cost of hospitalization of age groups, by surgeon volume

Adrenalectomy performed by a high-volume surgeon was associated with improved outcomes for older patients as well as for those ≤60 years. Among patients aged 61–70 years, complication rates were 19.4% lower, mean LOS was 27% shorter, and mean cost of hospitalization was 14.2% lower when adrenalectomy was performed by a high-volume surgeon (p = 0.15; p < 0.001; p = 0.002 respectively). Similarly, complication rates were 22% lower, mean LOS was 25% shorter, and mean cost of hospitalization was 13.2% lower when patients aged >70 years underwent adrenalectomy by a high-volume surgeon compared with a low-volume colleague (p = 0.17; p < 0.001; p = 0.043, respectively). The differences in complication rates were not statistically significant, likely due to low statistical power. To assess these findings under higher statistical power, the 61–70 and >70 years age groups were combined to create a larger >60 years age group (n = 2,231). The complication rate in this group was then analyzed by surgeon volume; we found a significant difference when patients >60 years underwent adrenalectomy by a high-volume surgeon compared with a low-volume surgeon (18.5 vs. 23.3%; p = 0.04).

Discussion

This study was designed to determine the impact of older age (>60 years) on clinical and economic outcomes after adrenalectomy. We demonstrated higher complication and mortality rates, longer LOS, and higher cost of hospitalization in older patients. We showed that adrenalectomy performed by high-volume surgeons resulted in improved clinical and economic outcomes among older patients, as well as among patients ≤60 years. After risk adjustment, older age remained independently associated with sustaining ≥1 complication, in-hospital mortality, and longer hospitalization after adrenalectomy.

There is a paucity of literature on age-stratified adrenalectomy outcomes. However, our results are consistent with adrenalectomy studies in which patient age was included as a covariate in analyses. In a study of endocrine surgery outcomes with regard to surgeon volume, Stavrakis et al. examined 576 adrenalectomy cases performed in the states of Florida and California.17 The authors found that patient age (modeled as a continuous variable) was an independent risk factor for sustaining a complication after adrenalectomy. Murphy et al. explored adrenalectomy outcomes over time and found a positive association between increasing age and adverse adrenalectomy outcomes.13 The authors demonstrated that age ≥65 years was independently associated with mortality after adrenalectomy.

The finding that older age is associated with worse surgical outcomes compared with those of younger patients is consistent with results of several studies. Sosa et al. showed that clinical and economic outcomes of patients 65 years and older undergoing thyroidectomies were considerably worse than for similar, younger patients.19 Rehospitalization of older (age >65 years) patients with thyroid cancer who underwent thyroidectomy has been shown to be prevalent and costly.20 A study involving more than 4,000 patients aged ≥50 years who underwent major noncardiac surgery also found a higher rate of major perioperative complications, mortality, and a longer length of stay among older patients.21 Another study involving more than 73,000 patients who underwent carotid endarterectomy found higher rates of cardiac, renal, and pulmonary complications in patients >60 years.22 Older age also was a risk factor for postoperative complications, mortality, increased LOS, and cost of hospitalization in more than 90,000 cholecystectomy patients.23 Reasons for increased risk of worse surgical outcomes among the elderly are complex; age appears to remain independently associated with these outcomes even after adjustment for comorbidities. In addition, there is evidence suggesting that perioperative care for the elderly, particularly regarding management of nutrition, polypharmacy, postoperative delirium, and deconditioning, is different and more difficult than for the nonelderly patient.4,24,25

We found that the frequency of postoperative adrenal insufficiency increased with age. Data examining the risk of adrenal insufficiency with respect to age are scant and contradictory.2628 Multiple studies also have indicated that diagnosing postoperative adrenal insufficiency can be challenging.2932 However, unexpected occurrence of adrenal insufficiency in patients who underwent adrenalectomy for a variety of adrenal tumors, including pheochromocytoma, malignant adrenal disease, and “non-functioning” adrenal tumors, has been reported in the literature; these occurrences often are attributed to the controversial entity of subclinical Cushing’s syndrome.29,30,3335 Exogenous glucocorticoid treatment is among the most common causes of impaired adrenal function, and older patients are more likely to be on long-term steroid treatment for chronic disease31,36 In our study, older patients were more to likely to have high comorbidity scores indicative of chronic disease; they also had higher rates of adrenal malignancy, bleeding, and infections. All of these factors could predispose older adrenalectomy patients to postoperative adrenal insufficiency.31,3638

Previous studies have demonstrated an association between surgeon volume and adrenalectomy outcomes. Using the state of Florida hospital discharge data, Gallagher et al. showed that low-surgeon adrenalectomy volume was associated with longer LOS.39 Also using state-level hospital discharge data, another study found that surgeon volume correlated inversely with LOS and cost after adrenalectomy in Florida and California.17 Park et al. demonstrated that low-surgeon volume was predictive of higher complication rates and longer LOS at a population level.16 Our results support these findings; indeed, low-surgeon volume was associated with higher complication rates, longer LOS, and higher costs in older patients as well as patients aged ≤60 years. Older patients were more likely to have surgery performed by a low-volume surgeon. Assuming surgeon volume is a proxy for experience, it seems that surgeon inexperience is associated with worse outcomes in older adrenalectomy patients.

The limitations of our study include those inherent to administrative databases, such as coding errors, but the NIS database is widely used and has been validated. NIS does not include information on extent of illness, such as stage of malignant disease, severity of illness or complications, do-not-resuscitate status, and laboratory data, such as adrenal biochemical levels. Readmissions are not included in the database; therefore, complication rates and mortality may have been underestimated. Although complexity of illness was assessed using the Charlson Comorbidity Index, frailty is a novel measure of physiologic reserve in the elderly that may not have been adequately estimated using this index.3 Due to the lack of a unique ICD-9-CM code specific to laparoscopic adrenalectomy, limited information about the impact of laparoscopy on our results can be drawn. However, the mean LOS of 3.7 days in this study is consistent with the average length of stay (approximately 3 days) for laparoscopic adrenalectomy reported in the literature.32,40 The strengths of this study are its large sample size, and the fact that it is the first to encompass national data, inclusive of a fifth of U.S. hospital admissions, to specifically analyze outcomes of adrenalectomy stratified by age.

Given the realities of an aging U.S. population and the increased detection of adrenal incidentalomas, more older patients will likely undergo adrenalectomy in the future. Our study demonstrates that the risk of adverse adrenalectomy outcomes is higher for older versus younger patients; nonetheless, these events occur infrequently overall. Most patients in our study had no complications and were discharged from the hospital in 3–4 days. Individualized assessment of surgical risks is necessary for all patients, regardless of age. Referral to a high-volume surgeon is a potentially modifiable factor, which may improve adrenalectomy outcomes for older patients. More research is needed to evaluate factors associated with the risk of adrenal insufficiency in older adrenalectomy patients and to examine long-term outcomes of these patients.

Acknowledgment

This study was supported in part by the Yale Medical Student Research Fellowship (Hadiza S. Kazaure).

Conflict of interest

The authors declare no conflicts of interest.

Copyright information

© Society of Surgical Oncology 2011