Abstract
ABSTRACT. Mortality rates after lower extremity amputation are extremely high among dialysis patients. However, the impact of milder degrees of renal insufficiency on death rates after lower extremity amputation has not been carefully examined. In this study, the authors used data from the Department of Veterans Affairs’ National Surgical Quality Improvement Program (NSQIP) to measure the association between renal dysfunction and 30-d mortality after nontraumatic amputation adjusted for confounders. The study population consisted of 16,994 patients undergoing their first NSQIP recorded amputation from January 1, 1994 through September 30, 2001. Thirty-five percent of all cohort patients had at least moderate renal insufficiency, and 52% of all postoperative deaths occurred in this group. Postoperative mortality was 9% in patients with moderate renal insufficiency, 15% in patients with severe renal insufficiency, and 16% in dialysis patients, compared with 6% in patients with normal or mildly reduced renal function. Renal insufficiency remained associated with death after adjustment for confounders (adjusted odds ratio [OR] 3.36, 95% confidence interval [CI] 2.75 to 4.10 [dialysis patients]; OR 2.54, CI 2.06 to 3.14 [severe renal insufficiency]; and OR 1.52, CI 1.32 to 1.76 [moderate renal insufficiency]). In conclusion, even moderate renal insufficiency is independently associated with postoperative death after lower extremity amputation. This finding highlights the need for a targeted approach to improving the care of patients with renal insufficiency undergoing lower extremity amputation.
Mortality rates after lower extremity amputation are extremely high among dialysis patients (1–5⇓⇓⇓⇓). However, the impact of milder degrees of renal insufficiency on death rates after lower extremity amputation has not been carefully examined. Several earlier studies have reported associations between renal insufficiency and death after amputation (6–8⇓⇓). However, in these studies, renal insufficiency was defined in a dichotomous fashion and was not the primary focus of the analyses. Therefore, it is not known whether there is a graded increase in risk of death after amputation with advancing degrees of renal insufficiency after adjustment for potential confounders.
The goals of the present analysis were to determine: (1) whether renal insufficiency was independently associated with death within 30 d after lower extremity amputation; and (2) whether there was a graded increase in risk of death after lower extremity amputation with advancing degrees of renal insufficiency. The existence of an independent association between renal insufficiency and postoperative death after amputation, and identification of a threshold level of renal function at which risk is increased, would help to determine whether a targeted approach to improving outcomes after lower extremity amputation in this population is warranted.
Materials and Methods
Database
The VA NSQIP is an ongoing quality management initiative for surgical care. Since its inception on January 1, 1994, NSQIP has prospectively collected data on most major surgeries occurring at VA medical centers (VAMC) across the country (9). At present, 123 VAMC participate in NSQIP. For each procedure included in NSQIP, surgical clinical nurse reviewers collect baseline demographic and clinical variables, and 30 d postoperative surgical outcomes at each site. Laboratory values are transmitted electronically from the VA’s decentralized hospital computer system (VistA) to the coordinating center at the Denver VA Medical Center and the University of Colorado Health Outcomes Program. The only major surgical procedures that are eligible for inclusion are those performed under general, spinal, or epidural anesthesia at the 123 VAMC participating in NSQIP.
Sample Selection
We searched the NSQIP database for all principal procedure nontraumatic lower extremity amputation Current Procedural Terminology (CPT) codes from January 1, 1994 through September 30, 2001 (CPT codes 28810 [metatarsal with single toe], 28820 [toe at the metatarsal-phalangeal joint], 28825 [toe at the interphalangeal joint], 28805 [transmetatarsal], 28800 [mid-tarsal (Chopart)], 27880–27882 [transtibial], 27888–27889 [ankle disarticulation], 27598 [knee disarticulation], 27590–27592 [transfemoral], 27295 [hip disarticulation], and 27290 [abdominopelvic]). We included only the chronologically first amputation procedure for each patient recorded in NSQIP during the study period to analyze all data at the individual level and to exclude as many secondary procedures as possible. For the current analysis, we divided amputation procedures into three general categories according to the anatomic level of the amputation: partial foot (toe and foot amputations), transtibial (including ankle disarticulation and knee disarticulation, in addition to transtibial amputations proper), and transfemoral or higher (if the amputation level was above the knee joint) (10).
Study Outcome
The outcome examined in this analysis was postoperative death within 30 d of amputation. All deaths occurring during this period were verified against the VA Beneficiary Identification and Records Locator Subsystem (BIRLS) death records (11).
Determination of Renal Function
A single preoperative serum creatinine measurement was available for most patients in the NSQIP database. In addition, nurse reviewers (who have access to the medical record) recorded whether patients experienced acute renal failure (predefined by NSQIP as a rapidly rising creatinine above 3 mg/dl and oliguria), or were on dialysis before surgery. For the present analysis, GFR was calculated for each patient using the abbreviated Modification of Diet in Renal Disease (MDRD) formula (12), which predicts GFR on the basis of serum creatinine, age, race, and gender. We excluded from the analysis all patients who had preoperative acute renal failure, as it was defined by NSQIP, to include only patients likely to have chronic renal insufficiency.
Patients were assigned to one of the four following groups according to their level of renal function: (1) normal or mildly reduced renal function (estimated GFR ≥60 cc/min per 1.73 m2); (2) moderate renal insufficiency (estimated GFR 30–59 cc/min per 1.73 m2); (3) severe renal insufficiency (estimated GFR < 30 cc/min per 1.73 m2); and (4) dialysis-dependent renal failure. This classification is similar to the definition recommended by the National Kidney Foundation’s Dialysis Outcome Quality Initiative (NKF-DOQI) guidelines for chronic kidney disease (13) with one exception. To draw comparisons between dialysis patients and nondialysis patients (a clinically relevant distinction), the small number (n = 157) of patients with an estimated GFR < 15cc/min per 1.73 m2 but not on dialysis were included in the severe renal insufficiency group (estimated GFR < 30cc/min per 1.73 m2); and not with dialysis patients in a “kidney failure” category as suggested by NKF-DOQI guidelines.
Measurement of Confounding Variables
Potential confounders included patient demographic, clinical, preoperative, and operative characteristics. Demographic characteristics included age, race (African American versus non-African American), and gender. Clinical characteristics included history of diabetes, history of congestive heart failure (CHF) during the month before surgery, history of stroke with residual neurological deficit, history of severe chronic obstructive pulmonary disease (COPD), smoking within 1 yr of admission, and current alcohol use (defined as more than two drinks per day during the 2-wk period before admission).
Patient preoperative condition included the presence of preoperative wound infection, sepsis within 48 h before surgery, preoperative white blood cell count, hematocrit, albumin, and impaired sensorium (defined as mental status changes or delirium in the context of the current illness). Operative characteristics included whether the procedure was performed on an emergency basis, whether it was performed under general anesthesia (versus spinal), whether the wound was contaminated, and the anatomic location of the amputation (partial foot versus transtibial versus transfemoral or higher).
Before 1996, NSQIP collected more extensive data on cardiovascular comorbidities (history of myocardial infarction and use of anti-hypertensive medications) that were not collected after this time. Data on these comorbities were therefore only available for a subset of patients in the sample (n = 2977). A sensitivity analysis was conducted in this patient subset to determine whether an association of renal insufficiency with death after amputation was still present after additional adjustment for these covariates.
Analysis of Postoperative Complications
To explore the possible causal pathway for increased mortality associated with renal insufficiency, we examined the incidence of major postoperative complications occurring within 30 d of surgery that were associated with death within 30 d by level of renal function. These included cardiac arrest requiring CPR, myocardial infarction (MI), unplanned intubation for respiratory/cardiac failure, prolonged intubation (>48 h), pulmonary embolism, postoperative sepsis, pneumonia, bleeding requiring >4 units of blood within the first 72 h after surgery, stroke, coma >24 h, urinary infection, and acute renal failure requiring dialysis. In addition, we measured the association of renal function with the need for at least one re-amputation at the same or more proximal anatomic level occurring between 1 and 30 d after the initial procedure. We measured this association because repeat amputation procedures may contribute to postoperative mortality. Re-amputations among cohort patients were identified by searching the VA administrative database (Patient Treatment File) for amputation procedures at the same or more proximal level occurring within 30 d of the original amputation procedure. We adopted this strategy to capture all repeat amputation procedures including those not recorded in NSQIP. Amputations in the Patient Treatment File were identified by International Classification of Diseases (ICD-9) procedure code search (ICD-9 84.11 through 84.19).
Statistical Analyses
Differences in patient characteristics by level of renal function were compared by χ2 test, using the group with normal or mildly reduced renal function (≥60 cc/min per 1.73 m2) as the reference category. Continuous variables were dichotomized as follows: age was dichotomized at the median value (68 yr), and other variables were dichotomized at clinically meaningful cutoffs (i.e., albumin ≤3 g/dl versus >3 g/dl; hematocrit ≤30% and >30%; and white blood cell count ≤12 and >12 cells/μl.)
We used logistic regression analysis to measure the association of renal function level with death within 30 d of lower extremity amputation, in both univariate and multivariate analyses. The multivariable model was adjusted for a random center effect to accommodate the potentially correlated nature of procedure data from individual surgical centers (14). All variables associated with both the outcome (death within 30 d) and with the exposure (renal insufficiency) in univariate logistic regression analysis were included in the multivariable model. For each analysis, patients with normal or mildly reduced renal function served as the reference category against which all others were compared. Observations missing data for a given variable were coded as “missing” and a missing indicator variable was included in the analysis. For any given observation, this variable was coded as 0 if data were present for the variable in question and 1 if data were missing. This strategy allowed all observations to be used in estimating the effects of variables that were not missing. Inclusion of the missing indicator variable ensured that when a separate missing category was generated, it would have no influence on the variables’ estimated effects.
To further examine the precise timing of postoperative death in each group, Kaplan-Meier curves were generated to examine, graphically, time to death by level of preoperative renal function over the 30-d postoperative period. We measured the incidence of major operative complications that were associated with death across renal function categories to explore variables that we postulated might be part of the causal pathway for death after amputation. The incidence of each complication for each renal insufficiency category was compared with the reference category of patients with normal or mildly reduced renal function using a χ2 test. We repeated the multivariable analysis (described above), additionally adjusting for postoperative complications, to examine whether an increased incidence of postoperative complications in patients with renal insufficiency accounted for the increased postoperative mortality seen in this group. Included in this analysis were all postoperative complications that were associated with death within 30 d of amputation and for which there was a statistically significant difference in incidence in at least one renal insufficiency group compared to the reference category.
Primary analyses were conducted using estimated GFR categories calculated with the abbreviated MDRD equation. Because the abbreviated MDRD formula has not been validated in a veteran population, we repeated the analysis using serum creatinine cutpoints for moderate and severe renal insufficiency of 1.57 mg/dl and 2.19 mg/dl for men, and 1.17 mg/dl and 1.77 mg/dl for women, respectively, as described by Couchoud et al. (15). As for the primary analysis, dialysis patients were included as a separate category. In addition, a sensitivity analysis was conducted to confirm the presence of an association between renal insufficiency and death after lower extremity amputation after adjustment for history of myocardial infarction and use of anti-hypertensive medications (available only before 1996), in addition to other potential confounders tested in the entire sample of patients. All analyses were conducted using Stata Statistical Software (Stata Corporation, College Station, TX).
Results
During the study period, the NSQIP recorded a total of 23,832 nontraumatic lower extremity amputations in 18,519 unique patients performed at 123 VAMC. Only the first recorded procedure for each patient was included in the analysis. A total of 497 patients were excluded because they had preoperative acute renal failure as defined by NSQIP. Another 1028 patients were excluded because they lacked data on renal function yielding an analytic sample of 16,994 first amputations recorded in NSQIP during the study period (3034 partial foot amputations, 7911 transtibial amputations, and 6049 amputations at the transfemoral level or higher).
While the majority of patients (n = 11,051) undergoing lower extremity amputations recorded in NSQIP had normal or mildly reduced renal function (GFR ≥60 cc/min per 1.73 m2), a substantial percentage (35%) had some renal insufficiency. This group consisted of 3902 patients with moderate renal insufficiency (GFR 30–59 cc/min per 1.73 m2), 931 patients with severe renal insufficiency (<30 cc/min per 1.73 m2), and 1110 dialysis-dependent patients. Mean serum creatinine values were 0.9 ± 0.2 mg/dl for patients with normal or mildly reduced renal function, 1.7 ± 0.3 mg/dl for patients with moderate renal insufficiency, 3.6 ± 1.6 mg/dl for patients with severe renal insufficiency, and 6.1 ± 2.7 mg/dl for patients undergoing dialysis.
Table 1 shows that data were complete for most variables. Several variables indicated in the table were missing <1% of values. However, data on serum albumin and the presence of preoperative sepsis were missing for 29% and 18% of patients, respectively. There were substantial differences in baseline characteristics by level of renal function. Compared with patients with normal renal function, patients in all renal insufficiency groups were more likely to have diabetes, congestive heart failure, anemia, hypoalbuminemia, leukocytosis, or sepsis before amputation and less likely to use alcohol or to smoke. They were also more likely to have transtibial (versus other) amputations and to have contaminated wounds.
Table 1. Characteristics of patients undergoing a first amputation procedure recorded by the VA National Surgical Quality Improvement Program startified by level of renal function, 1994–2001
A total of 1318 postoperative deaths (within 30 d of the principal procedure) occurred during the study period. Over half of these were in patients with renal insufficiency. Death rates for all renal insufficiency groups were higher than for patients with normal renal function (Table 2). In unadjusted logistic regression analysis, there was a strong association between level of renal function and odds of postoperative death (Table 2). Adjustment for demographic characteristics, comorbid conditions, preoperative condition, and operative characteristics slightly attenuated the association of death with moderate and severe renal insufficiency, but actually increased the magnitude of the association with dialysis-dependent renal failure. Overall, adjusted and unadjusted odds ratios (OR) are quite similar for each category of renal function. In addition, a statistically significant center effect was noted in the multivariable model (P = 0.001). Among the subgroup of 2977 patients for whom additional cardiovascular comorbidity data were available (history of myocardial infarction and use of anti-hypertensive medications), additional adjustment for these comorbidities did not substantially alter the association of renal insufficiency with death after amputation (patients with moderate renal insufficiency, OR 1.29, CI 0.93 to 1.79; patients with severe renal insufficiency, OR 2.81 CI 1.77 to 4.45; and dialysis-dependent patients, OR 3.84, CI 2.24 to 6.57).
Table 2. Unadjusted and adjusted logistic regression analyses of the association of level of renal function with postoperative death among patients undergoing a first amputation recorded by the VA National Surgical Quality Improvement Program, 1994–2001a
To determine the stage at which deaths occurred during the postoperative period, we conducted a survival analysis using time to death as the outcome and censoring all survivors at 30 d after the principal procedure. Kaplan-Meier curves for each renal insufficiency category are presented in Figure 1. Deaths were spread evenly throughout the postoperative period for patients with normal renal function and moderate and dialysis-dependent renal insufficiency. However, patients with severe renal insufficiency not on dialysis had a higher mortality rate than all other groups for most of the postoperative period (first 25 d) that leveled off by 30 d.
Figure 1. Kaplan-Meier curves by level of renal function for patients undergoing first lower extremity amputation recorded in the VA National Surgical Quality Improvement Program, 1994–2001.
Table 3 shows, by level of renal function, the incidence of postoperative complications occurring within 30 d of the initial amputation procedure. Of note, patients with renal insufficiency had a higher incidence of second amputations occurring at the same or a higher anatomic level within 30 postoperative days, and the incidence of this complication increased across renal insufficiency categories. In addition, most cardiopulmonary complications were more common in patients with renal insufficiency than in those with normal renal function. When the multivariable model (described above) was additionally adjusted for postoperative complications, the odds of postoperative death for patients with renal insufficiency was only slightly attenuated (Table 4). Results for all univariate and multivariate analyses (described above) were not substantially different when we classified patients’ renal function by creatinine cutpoints rather than by estimated GFR.
Table 3. Incidence of postoperative complications (within 30 d) among patients undergoing a first lower extremity amputation recorded by the VA National Surgical Quality Improvement Program, 1994–2001
Table 4. Multivariable logistic regression analysis of the association of level of renal function with postoperative death additionally adjusted for postoperative complications among patients undergoing a first amputation recorded by the VA National Surgical Quality Improvement Program, 1994–2001
Discussion
Patients with renal insufficiency accounted for over one third of all lower limb amputations and over half of the postoperative deaths occurring within 30 d of amputation in this cohort. Furthermore, over half of all postoperative deaths in patients with renal insufficiency occurred in those with only moderate renal insufficiency (mean serum creatinine, 1.7 mg/dl; range 1 to 2.9 mg/dl). Strikingly, adjustment for a wide range of potential confounders made little impact on the strength of the association of renal insufficiency with death within 30 d of amputation. These findings support a strong independent association between renal insufficiency and postoperative death after amputation and are consistent with the wider literature identifying moderate renal insufficiency as a risk factor for death after both cardiac (16) and non-cardiac (17–21⇓⇓⇓⇓) surgery.
We found that most postoperative complications associated with death (including cardiovascular complications) were more common in patients with renal insufficiency than in those with normal renal function, including the incidence of at least one repeat lower extremity amputation at the same or at a higher anatomic level during the 30-d postoperative period. However, when the analysis was adjusted for these complications, odds ratios for the association of each category of renal insufficiency with postoperative death were only slightly attenuated. This finding suggests that the more frequent occurrence of postoperative complications in patients with renal insufficiency does not fully explain the higher death rate in this group. Rather, the increased incidence of postoperative death among patients with renal insufficiency appears to be out of proportion to the increased incidence of recorded postoperative cardiovascular complications in this group.
Possible explanations for the increased risk of postoperative death in patients with renal insufficiency include increased severity of complications, increased severity of underlying comorbid conditions and suboptimal management, and/or underdiagnosis of complications such as acute myocardial infarction as has been seen in other settings (22,23⇓). In addition, the high prevalence of preoperative sepsis and leucocytosis at the time of initial amputation, the more frequent need for emergency procedures (with the exception of dialysis patients), and the higher incidence of repeat amputations in patients with renal insufficiency raise the question of whether operative intervention is occurring relatively late in the course of lower extremity disease in this patient group, perhaps resulting in systemic illness. Detailed studies of the postoperative course of patients with renal insufficiency undergoing lower extremity amputation may help to identify reasons for increased mortality and suggest pathways for improving outcomes in this group.
The higher mortality after lower extremity amputation among patients with renal insufficiency also highlights the importance of strategies to prevent amputation in this population. Research among persons with diabetes in the general population has shown that early recognition and management of risk factors foot ulceration and amputations can prevent or delay the onset of adverse outcomes (24). Several simple interventional approaches to preventing amputation and ulceration have been reported in the literature and include patient and provider education and regular foot examinations (25,26⇓). Importantly, at least one recent study has shown that implementation of a preventive foot care program is feasible in the dialysis unit setting (27). However, despite the relatively high risk and potentially devastating consequences of amputation among end-stage renal disease patients, few efforts are currently made to provide preventive foot care to this population and screen for risk of future amputation (28). For example, the comprehensive guidelines provided by the National Kidney Foundation’s Dialysis Outcome Quality Initiative (NKF-DOQI) that address many important issues in the acute and preventive care of patients with chronic kidney disease do not include foot care guidelines (http://www.kidney.org/professionals/doqi). Given the high postoperative mortality associated with amputation in patients with renal insufficiency, more emphasis should be placed on developing preventive foot care programs for both dialysis patients and for patients with milder degrees of renal insufficiency.
The major limitation of this analysis is that the results may not be generalizeable to nonveteran populations. In addition, we used estimated GFR rather than directly measured GFR, which was not available. The NSQIP database does not contain patient weight, so we were unable to use the Cockroft-Gault equation to estimate GFR. Instead, we opted to use the abbreviated MDRD equation, which is adjusted for patient race, gender, and age. However, the transportability of the abbreviated MDRD equation to a veteran population has not been tested, and creatinine measurements collected by NSQIP have not been calibrated (29). This may have resulted in an overestimation of the number of patients with renal insufficiency, but it is doubtful that misclassification of renal function would have strengthened the associations described here. Also, we are reassured by the fact that the associations reported here did not change with the use of serum creatinine cut points instead of estimated GFR (30).
In addition, we cannot be certain that all patients with renal insufficiency met the definition of chronic kidney disease (decreased GFR over at least a 3-mo period) or that all dialysis patients in this study were on chronic (versus acute) dialysis. Thus, we cannot be certain that our results can be generalized to patients with chronic kidney disease. However, we excluded from our analysis all patients identified as having preoperative acute renal insufficiency by nurse reviewers, and our results are consistent with prior studies in chronic dialysis patients.
Conclusions
Thirty-five percent of first lower extremity amputations and over 50% of postoperative deaths recorded in NSQIP during the study period occurred in patients with renal insufficiency. Even moderate renal insufficiency was independently associated with postoperative death. For most of the postoperative period, mortality was actually higher for patients with severe renal insufficiency not on dialysis than for those on dialysis at the time of amputation. In addition, patients with renal insufficiency were more likely to undergo repeat amputation at the same level or higher within 30 d of their initial NSQIP amputation. While reamputation along with most major postoperative complications was more common among patients with renal insufficiency than among those with normal renal function, the increased incidence of these complications alone did not explain the increased incidence of death in this group. These findings suggest the need for careful examination of preoperative decision-making and perioperative management for patients with renal insufficiency undergoing lower extremity amputation, and a targeted approach to improving outcomes after amputation in this group. In addition, our findings underline the importance of more aggressive efforts to prevent lower extremity amputation in this population.
Acknowledgments
We thank Bharat Thakkar, Dr. Barbara Krasnicka, and Robbin Denwood for their programming and informational support, and Drs. Seth Landefeld, Kenneth Covinsky, Louise Walter, Kala Mehta, Alex Li, Kenneth Chuang, Deborah Barnes, and other members of the San Francisco VAMC Geriatrics Department for providing helpful feedback on an earlier draft of this manuscript. Dr. O’Hare is supported by a Health Services Research Career Development Award from the Department of Veterans Affairs. Part of this research was performed while Dr. O’Hare was supported by a fellowship-training grant from the National Kidney Foundation with matching support from the National Kidney Foundation of Northern California.
- © 2004 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Diabetes, Lower-Extremity Amputation, and Death
- Impact of Chronic Kidney Disease on Survival After Amputation in Individuals With Diabetes
- Peripheral Vascular Disease-Related Procedures in Dialysis Patients: Predictors and Prognosis
- Renal Insufficiency and the Risk of Lower Extremity Peripheral Arterial Disease: Results from the Heart and Estrogen/Progestin Replacement Study (HERS)