- Email: [email protected]
Relation between gender and vascular access complications in hemodialysis patients
Relation Between Gender and Vascular Access Complications in Hemodialysis Patients Brad C. Astor, MS, MPH, Josef Coresh, MD, PhD, Neil R. Powe, MD, MPH, MBA, Joseph A. Eustace, MB, MRCPI, MHS, and Michael J. Klag, MD, MPH ● Native arteriovenous (AV) fistulae for hemodialysis vascular access are believed to be associated with fewer complications than synthetic polytetrafluoroethylene (PTFE) grafts. We conducted a study among patients in the Dialysis Morbidity and Mortality Study to compare risk factors for complications of AV fistulae and PTFE grafts in men and women and to examine the effect of age on vascular access complications. We analyzed data from 833 incident patients with end-stage renal disease who had a PTFE graft (n ⴝ 621) or AV fistula (n ⴝ 212) in use 1 month after starting hemodialysis therapy. Follow-up using inpatient and outpatient Medicare administrative data identified a 1.8-times greater risk for a subsequent vascular access procedure for PTFE grafts (0.71 procedures/accessyear) than for AV fistulae (0.39 procedures/access-year). Men with grafts and women with grafts or fistulae had a greater risk for a first subsequent access procedure than did men with fistulae (0.79, 0.65, and 0.59 versus 0.33 procedures/access-year, respectively). After adjustment for age, race, presence of diabetes mellitus, and history of smoking, peripheral vascular disease, and cardiovascular disease, use of a PTFE graft compared with an AV fistula was associated with a greater risk for a first subsequent procedure in men (relative hazard, 2.2; 95% confidence interval [CI], 1.6 to 2.9), but not in women (relative hazard, 1.0; 95% CI, 0.7 to 1.4). The excess risk associated with a PTFE graft compared with an AV fistula was limited to men in the lower three quartiles of age (ie, <72 years). These data raise concern that the potential benefits of AV fistulae over PTFE grafts are not realized in women and older men. A better understanding of the determinants of successful access maturation and maintenance in these groups is needed. © 2000 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis (HD); vascular access; arteriovenous (AV) fistula; polytetrafluoroethylene (PTFE) graft; gender; age; vascular access complications; thrombosis.
V
ASCULAR ACCESS complications are the leading cause of morbidity in chronic hemodialysis patients, accounting for 15% of the hospitalizations in this population.1 Medicare From the Departments of Epidemiology, Biostatistics, and Health Policy and Management, The Johns Hopkins School of Hygiene and Public Health; Welch Center for Prevention, Epidemiology, and Clinical Research, The Johns Hopkins University; and the Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD. Received February 16, 2000; accepted in revised form June 30, 2000. Supported in part by grant no. T32HL07024-23 from the National Heart, Lung, and Blood Institute (B.C.A.); grants no. R29DK48362 (J.C.), K24DK02643 (N.R.P.), T32DK07732 (M.J.K.), and K24DK02856 (M.J.K.) from the National Institute of Diabetes and Digestive and Kidney Diseases; and grant no. R01-HS-08365 from the Agency for Health Care Research and Quality. The data reported here have been supplied by the US Renal Data System. The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy or interpretation of the US government. Address reprint requests to Josef Coresh, MD, PhD, 2024 E Monument St, Ste 2-600, Baltimore, MD 21205. E-mail: [email protected] © 2000 by the National Kidney Foundation, Inc. 0272-6386/00/3606-0005$3.00/0 doi:10.1053/ajkd.2000.19816 1126
expenditures for vascular access–related procedures were nearly $1 billion in 1994, representing between 14% and 17% of the total spending for hemodialysis patients.2 Most studies have found native arteriovenous (AV) fistulae to be associated with improved access survival and fewer complications than prosthetic polytetrafluoroethylene (PTFE) grafts.3-9 However, the proportion of patients starting chronic hemodialysis therapy in the United States who receive an AV fistula has declined in the last decade.9,10 In 1993, more than two thirds of the permanent accesses in use at 1 month after starting hemodialysis were PTFE grafts.11 The reasons for this trend are unclear but may be the increasing mean age at incidence of end-stage renal disease (ESRD) and the increasing proportion of patients with diabetes mellitus as the cause of ESRD.10,12 Age and diabetes mellitus, as well as female sex, have been associated with a greater likelihood of having a PTFE graft than an AV fistula in place 30 days after starting hemodialysis.9 Several factors have been postulated as predictive of vascular access complications, including having a PTFE graft rather than an AV fistula, black race, advanced age, diabetes mellitus, and a history of peripheral vascular disease.13,14
American Journal of Kidney Diseases, Vol 36, No 6 (December), 2000: pp 1126-1134
GENDER AND VASCULAR ACCESS EVENTS
Woods et al14 recently reported that the risk for an access complication requiring intervention increased significantly with increasing age for AV fistulae but not PTFE grafts. They found that AV fistulae and PTFE grafts had similar rates of complications in patients aged older than 65 years. Published data are conflicting on the effect of gender on rates of access-related complications. Several studies1,15,16 found greater rates of complications in women than men, whereas many others found no effect.8,14,17-19 Few of these studies examined the effect of gender among patients with AV fistulae separately from the effect among patients with PTFE grafts, and none has included a direct comparison of AV fistulae to PTFE grafts in men and women separately.16,17,20 The current study examines vascular access complications in a large cohort of incident hemodialysis patients who had a permanent access in use 1 month after starting hemodialysis therapy. The study compares complications in men and women and tests whether the benefit of AV fistulae over PTFE grafts is smaller in older individuals. In addition, the study examines the effects of clinical variables, including access type, on vascular access complications. METHODS
Study Design and Population We conducted a national longitudinal study of vascular access complications in incident hemodialysis patients. The study population was drawn from the 1,997 patients included in the vascular access component of Wave 1 of the Dialysis Mortality and Morbidity Study (DMMS) from the US Renal Data System (USRDS).21 The vascular access component of Wave 1 included all patients who started hemodialysis therapy during December 1993 at 550 randomly selected hemodialysis units and a random sample of patients from the same units who had started hemodialysis therapy earlier in 1993 and were undergoing hemodialysis on December 31, 1993. Our study included only those patients with a PTFE graft or AV fistula in use 1 month after starting hemodialysis therapy (n ⫽ 833). Patients were excluded for the following reasons: type of access in use was a temporary catheter (n ⫽ 705), bovine graft (n ⫽ 34), or missing (n ⫽ 91); the date of first chronic hemodialysis session was missing or was not during 1993 (n ⫽ 115); there was a lapse in all claims for greater than 60 days during the follow-up period (n ⫽ 131); or an access-related claim was reported before 1 month after starting hemodialysis therapy (n ⫽ 6). Because smoking was studied as a risk factor for subsequent vascular access procedures, an additional 82 patients were excluded for whom smoking information was missing.
1127
Data Collection Baseline data were obtained from the DMMS Standard Analysis File from the USRDS.21 Data were abstracted from patient medical records by dialysis facility personnel from March through July 1995. Data included information on patient demographics, comorbid conditions present in the last 10 years, psychosocial factors, laboratory values, dialysis treatments, type of vascular access in use 1 month after starting chronic hemodialysis therapy, and the date this access was placed. Race was categorized as black or other. Smoking was defined as any or no history of smoking. Additional analyses considered current smoking and previous smoking separately. Patients were considered to have a positive history of cardiovascular disease if they had undergone coronary angioplasty in the past or had been previously diagnosed with any of the following: angina, myocardial infarction, coronary heart disease, or coronary artery disease. Follow-up data for up to 4 years were obtained from detailed USRDS records that contained information on Medicare claims for hospital inpatient and outpatient services through December 31, 1996. Vascular access procedure codes from the International Classification of Diseases, Ninth Revision22 and the Physicians’ Current Procedural Terminology, Fourth Edition23 were included as outcomes. These codes correspond to creation of a new access, insertion of a venous catheter, and thrombectomy, revision, or removal of an access. The first date listed for a claim was used as the date of the access procedure.
Statistical Analysis Follow-up time was calculated starting 1 month after the initiation of chronic hemodialysis therapy until the date of the first access-related procedure recorded (hereafter referred to as the first subsequent procedure). Patients were censored at the time of death, renal transplantation, switch to peritoneal dialysis, or, if no event had occurred, on December 31, 1996 (the last date of follow-up). Time to the first subsequent procedure was analyzed using Kaplan-Meier estimates24 for the entire cohort and after stratification by access type, sex, and age. Cox proportional hazards models25 were used to examine the independent effect of sex on access complications after adjustment for demographic (age and race) and clinical (presence of diabetes mellitus or history of smoking, peripheral vascular disease, and cardiovascular disease) characteristics. The effect of age on the risk for a first subsequent procedure was further studied by dividing the male and female populations into four groups, each based on approximate quartiles of age. Access type– specific incidence rates and adjusted relative hazards associated with a PTFE graft versus an AV fistula were determined for each age group. Statistical analyses were performed using SAS (SAS Institute, Cary, NC)26 and Stata (StataCorp, College Station, TX).27
RESULTS
Study Population Characteristics The characteristics of patients who met the inclusion criteria for this study were compared
1128
ASTOR ET AL
with the 1,164 patients in Wave I who were excluded. The mean age of the study patients was 1.4 years older than the excluded patients (P ⬍ 0.005), but the groups did not differ in terms of sex, race, presence of diabetes mellitus, history of peripheral vascular disease, or history of cardiovascular disease. The study population consisted of 833 patients with a PTFE graft (n ⫽ 621) or AV fistula (n ⫽ 212) in use 1 month after starting hemodialysis therapy. Mean age at the time of first chronic hemodialysis session was 63.0 ⫾ 14.2 (SD) years. Nearly half the patients (406 patients) were women, and one third (281 patients) were black (Table 1). Patients receiving a PTFE graft were 4 years older (P ⬍ 0.001) and more likely to be women (P ⬍ 0.001), be black (P ⬍ 0.01), and have diabetes (P ⬍ 0.01). Patients with a PTFE graft were also more likely to have a history of peripheral vascular disease and cardiovascular disease, but these differences were not statistically significant (P ⫽ 0.14 and 0.06, respectively). Vascular Access Outcomes Follow-up through 1996 identified 552 first access–related procedures occurring more than 1 month after initiation of chronic hemodialysis therapy (ie, first subsequent procedures), consisting of the creation of 90 new accesses (16.3%), 59 venous catheter insertions (10.7%), and 403 Table 1. Characteristics by Access Type of 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis
Patient Characteristic
Age quartile 1 (18-54 y) 2 (55-64 y) 3 (65-72 y) 4 (ⱖ73 y) Women Black Diabetes mellitus Ever smoked History of peripheral vascular disease History of cardiovascular disease NOTE.
AV Fistula (n ⫽ 212)
PTFE Graft (n ⫽ 621)
69 (32.6) 42 (19.8) 54 (25.5) 47 (22.2) 65 (30.1) 56 (26.4) 90 (42.5) 99 (46.7) 44 (20.8)
139 (22.4) 141 (22.7) 174 (28.0) 167 (26.9) 341 (54.9) 225 (36.2) 330 (53.1) 277 (44.6) 160 (25.8)
77 (36.3)
272 (43.8)
P
⬍0.001 ⬍0.001 ⬍0.01 ⬍0.01 0.60 0.14
Values expressed as number (percent).
0.06
other procedures (73.0%; eg, thrombectomies, revisions, and removals). Overall median follow-up was 13.0 months. The incidence rate of first subsequent procedures was 0.61 procedures/ access-year. Of the 552 initial procedures identified, 442 occurred in PTFE grafts (incidence rate, 0.71 procedures/access-year) and 110 in AV fistulae (incidence rate, 0.39 procedures/accessyear). Univariate Associations With Initial Vascular Access Procedures AV fistulae had better procedure-free survival in men than women (P ⬍ 0.01; Fig 1). The median procedure-free survival among men with an AV fistula was 929 days compared with 329 days among women with an AV fistula. This sex difference was apparent (P ⬍ 0.02 by log rank) within the first 30 days of follow-up, by which time an access-related procedure had been performed in 14% of the women with an AV fistula compared with 5% of the men with an AV fistula. PTFE grafts had slightly better procedure-free survival in women than men (P ⬍ 0.05), with median procedure-free survival times of 332 and 267 days, respectively. Multivariate Associations With Initial Vascular Access Procedures Use of a PTFE graft compared with an AV fistula was associated with more than a twofold increase in the risk for an access-related procedure in multivariate proportional hazards analysis in men (relative hazard, 2.15; 95% confidence interval [CI], 1.62 to 2.85), but not in women (relative hazard, 1.03; 95% CI, 0.73 to 1.44), after adjustment for age, race, presence of diabetes, and history of smoking, peripheral vascular disease, and cardiovascular disease (Table 2). The association of peripheral vascular disease with initial access procedures reached statistical significance in men (relative hazard, 1.38; 95% CI, 1.09 to 1.67), but not in women (relative hazard, 1.19; 95% CI, 0.90 to 1.49). Conversely, the association of any history of smoking with initial access procedures reached statistical significance in women (relative hazard, 1.33; 95% CI, 1.07 to 1.59), but not in men (relative hazard, 1.18; 95% CI, 0.94 to 1.43). In a combined analysis of men and women, the presence of peripheral vascular disease (relative haz-
GENDER AND VASCULAR ACCESS EVENTS
1129
Fig 1. Kaplan-Meier estimates of freedom from first subsequent vascular access procedure by access type and sex for 833 incident hemodialysis patients with a permanent access in use 1 month after starting chronic hemodialysis therapy.
ard, 1.24; 95% CI, 1.01 to 1.52) and smoking (relative hazard, 1.25; 95% CI, 1.05 to 1.49) were associated with an increased risk for procedures after adjustment for the other factors in Table 2. Similar results were obtained from models considering current (n ⫽ 160) and previous (n ⫽ 216) smoking separately. Age, race, presence of diabetes mellitus, and history of cardiovascular disease did not predict first subsequent procedures in either sex or overall. In analyses
stratified by access type, female sex (versus male sex) was associated with a relative hazard of 1.58 (95% CI, 1.05 to 2.35) in patients with an AV fistula compared with a relative hazard of 0.90 (95% CI, 0.73 to 1.09) in patients with a PTFE graft. Modification of PTFE Graft Complication Risk by Age in Men and Women Among men, use of a PTFE graft compared with an AV fistula was associated in unadjusted
Table 2. Adjusted Relative Hazard of First Subsequent Vascular Access Procedure Associated With Patient Characteristics for 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis by Sex Men (n ⫽ 427)
P
Adjusted Relative Hazard
95% Confidence Interval
— 1.62 ⫺2.85 0.92 ⫺1.11 0.95 ⫺1.65 0.78 ⫺1.28 0.93 ⫺1.51 1.03 ⫺1.84
— ⬍0.001
1.0 1.03
0.85
1.01
0.11
1.09
0.99
0.99
0.18
1.33
0.03
1.19
— 0.73 ⫺1.44 0.92 ⫺1.11 0.86 ⫺1.39 0.76 ⫺1.28 1.03 ⫺1.72 0.89 ⫺1.60
0.63 ⫺1.09
0.17
1.21
Adjusted Relative Hazard
95% Confidence Interval
AV fistula PTFE graft
1.0 2.15
Age (per 10 y)
1.01
Black (v other race)
1.25
Presence of diabetes mellitus (v absence) Ever smoked (v never)
1.00
Patient Characteristic
Positive history of peripheral vascular disease (v negative history) Positive history of cardiovascular disease (v negative history)
1.18 1.38
0.83
Women (n ⫽ 406)
0.94 ⫺1.55
P
— 0.88 0.85 0.48 0.94 0.03 0.23
0.15
1130
ASTOR ET AL
analysis with a greater risk for first subsequent procedures in patients in the lower three quartiles of age (ie, ⬍73 years; P ⬍ 0.01 by log rank), but not in older patients (P ⫽ 0.67; Table 3). Among women, there was no statistically significant difference in risk among patients with PTFE grafts and those with AV fistulae in any age quartile. Multivariate proportional hazards analysis stratified by age quartile produced similar results after adjusting for race, presence of diabetes, smoking history, history of peripheral vascular disease, and history of cardiovascular disease. The relative hazard of first subsequent procedures associated with having a PTFE graft (compared with an AV fistula) varied with age quartile in men; 2.78 (95% CI, 1.66 to 4.65) in the lowest age quartile, 2.16 (95% CI, 1.10 to 4.26) in the second age quartile, 2.60 (95% CI, 1.26 to 4.46) in the third age quartile, and 1.02 (95% CI, 0.58 to 1.82) in the highest age quartile. Among women, the relative hazard was not significantly different from 1.0 in any group or overall and showed an inconsistent pattern across age groups. The number of AV fistulae in women (n ⫽ 65) limits the precision of these estimates. DISCUSSION
Previous studies have shown AV fistulae to have superior outcomes for hemodialysis patients compared with prosthetic grafts.3,5,8,18 The present study shows a significant benefit associ-
ated with use of an AV fistula in men, but not in women. These results held after adjusting for age, race, presence of diabetes mellitus, smoking history, history of peripheral vascular disease, and history of cardiovascular disease. This sexspecific benefit of AV fistulae over PTFE grafts has not been reported previously. An association of female sex with a greater risk for accessrelated hospitalization has been reported in two large studies,1,15 although neither investigated the results by access type. This distinction is important because it has been shown consistently that women are more likely to receive a PTFE graft than men.9,14,18,19,28 Many other studies found no association between sex and vascular access complications, but these also did not distinguish between AV fistulae and PTFE grafts.8,14,18,19 The effect of sex on access-related complications has also been examined in studies limited to patients using an AV fistula4,16,17,20 or to patients using a PTFE graft.17 For example, Aman et al17 found no sex differences in access failure or complication rates in 91 patients with PTFE grafts. Conversely, Golledge et al16 recently reported results from 107 patients (73 men, 34 women) with a first AV fistula; they found lower access survival for women after 2 years of follow-up (P ⬍ 0.01). Prischl et al20 found no significant differences in access survival between 80 men and 43 women on hemodialysis therapy with a first AV fistula
Table 3. Incidence Rate Per Access-Year of First Subsequent Vascular Access Procedure for 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis by Access Type, Sex, and Age Quartile Age Quartile
Men AV fistula (n ⫽ 147) PTFE graft (n ⫽ 280) P Women AV fistula (n ⫽ 65) PTFE graft (n ⫽ 341) P NOTE.
1 (18-54 y)
2 (55-64 y)
3 (65-72 y)
4 (ⱖ73 y)
Overall
0.29 (0.19-0.43)
0.28 (0.16-0.50)
0.29 (0.18-0.48)
0.58 (0.36-0.94)
0.33 (0.26-0.42)
0.99 (0.77-1.28)
0.72 (0.53-0.98)
0.83 (0.64-1.08)
0.65 (0.49-0.86)
0.79 (0.69-0.91)
⬍0.001
⬍0.001
⬍0.001
0.67
⬍0.001
0.41 (0.21-0.83)
0.52 (0.29-0.95)
0.60 (0.32-1.11)
0.96 (0.55-1.69)
0.59 (0.43-0.80)
0.70 (0.53-0.92)
0.55 (0.42-0.71)
0.78 (0.62-0.98)
0.60 (0.46-0.78)
0.65 (0.57-0.74)
0.21
0.34
0.14
0.14
0.74
Values expressed as incidence rate (95% CI). P by log rank.
GENDER AND VASCULAR ACCESS EVENTS
(P ⫽ 0.24), although the survival curves show poorer access survival for women at 1 year. Although all these studies have yielded valuable results, the evidence that may guide clinical practice is limited. Sample sizes have been small, and patient populations with AV fistulae were derived from single European centers. Differences in access placement practices between Europe and the United States limited the extent to which results could be generalized to the United States. Our study addresses these shortcomings, and the incidence rates for first access–related procedures occurring after 1 month of hemodialysis therapy (0.71 procedures/access-year for PTFE grafts; 0.39 procedures/access-year for AV fistulae) observed in this study are similar to those reported elsewhere.5,14,18,29,30 The results of previous studies are consistent with our findings that AV fistula–related procedure rates are greater in women than men, but PTFE graft–related procedure rates are similar by sex. However, because our study included both access types and had a larger sample, we were able to directly compare AV fistulae with PTFE grafts among men and women separately. Feldman et al1 suggested that differences in vessel diameter may account for the increased access-related morbidity experienced by women in their study. In a report of surgical placement of AV fistulae, Reilly et al31 found that vein size was a significant predictor of subsequent fistula survival, although sex was not. A more recent European study also found that small vessel size predicted fistula failure in the first 3 months after surgery.32 The effect of vessel size may be of less significance in patients with PTFE grafts than AV fistulae. This sex difference in vessel diameter has also been suggested as a reason for the poorer outcomes observed in women after coronary artery bypass surgery.33 In several studies of coronary artery bypass grafting that included measurements of vessel size, this characteristic was found to be the most predictive of outcome, diminishing any effect of sex.34-36 Sex differences in hemostatic factors have also been suggested as an explanation for the increased rate of vascular access complications in women.15 Our data show that women with AV fistulae had significantly more early access-related proce-
1131
dures (ie, during the first month of the study, which corresponds to the second month of hemodialysis therapy) than did their male counterparts. This result has been observed in previous studies4,16 and suggests that lack of successful access maturation is a critical element in the greater rate of complications associated with AV fistulae in women. Our study confirms the findings of Woods et al14 in the USRDS Case Mix Adequacy Study population in that the benefit of AV fistulae over PTFE grafts was found only in younger patients. However, the benefit we observed was exclusive to men. Previous studies have investigated the association of age with access complications but did not consider the effects of access type or sex. Several previous reports failed to find an association between age and access complications,8,18,19 although others reported a significant effect of age appearing after 6 months of follow-up20 and in patients without diabetes with prosthetic grafts.12 A review of USRDS data reported a significantly greater rate of access-related hospitalizations among those aged older than 64 years,1 although access type was not available for these patients. Woods et al14 reported a relative hazard of 1.40 for access procedures associated with a 10-year older age for an AV fistula in patients aged younger than 65 years and no significant effect of age in older patients or in those with a PTFE graft. The adjusted relative hazard of a PTFE graft versus an AV fistula decreased in that study from 3.0 at the age of 40 years to 1.3 at the age of 65 years and older. Among men in the present study, use of a PTFE graft compared with an AV fistula was associated with significantly lower procedure-free access survival in the three younger age quartiles; however, in the oldest age quartile (ⱖ73 years), the benefit was smaller and not statistically significant. The present study also confirms an earlier report of a significant association between a history of peripheral vascular disease and subsequent access procedures.14 This association is not surprising because the condition of the native vasculature is likely a determinant of the adequacy of any access created. The pathophysiological characteristics of peripheral vascular disease and vascular access complications may also be similar. It has been postulated that periph-
1132
eral vascular disease may compromise blood flow in the access, producing a greater risk for thrombosis.37 Smoking history has been examined in a few studies of vascular access morbidity, with inconsistent results.18,31,38 In this study, current or previous smoking was associated with a relative hazard of 1.25 (95% CI, 1.05 to 1.49) for subsequent access events. Analyzed separately, the relative hazard of previous smoking was nearly identical to that of current smoking. Previous peripheral vascular damage in former and active smokers, leading to acute access thrombosis, may partially explain this observation. The results of the present study have several important implications. The finding that AV fistulae have no benefit in terms of procedure-free access survival in women and older men suggests that the preferred type of permanent vascular access for all patients may not be as straightforward as described in the National Kidney Foundation’s Dialysis Outcomes Quality Initiative clinical practice guidelines.30 Whether any long-term benefits, such as improved access survival after interventions, are gained by women and older men with AV fistulae needs to be investigated in prospective studies. Such studies ideally should include data on the creation and maintenance of permanent accesses before the initiation of chronic hemodialysis therapy to evaluate the entire benefit of creating a functioning AV fistula in these patients. The results of these studies may show that a different costbenefit analysis regarding the optimal choice of access type is necessary for patients with different characteristics. The results of the present study also have implications relating to screening for access dysfunction. Compared with men with an AV fistula, women with an AV fistula experienced significantly more interventions during the first 3 months of hemodialysis therapy in our study. This finding suggests that women may require more frequent screening for early fistula dysfunction and that this frequent screening should continue for at least 3 months after the initiation of hemodialysis therapy. This finding may also suggest that women require a longer time for fistula maturation before its first use. During the present study (1993 to 1996), it is probable that few of
ASTOR ET AL
the participating centers had consistent access screening programs. Thus, it is unlikely that the greater rates of access-related procedures observed in women were caused by access problems detected during routine screening. This study has several limitations. Despite the large size and representative nature of the Wave 1 patient population, some subgroups analyzed in this study include a relatively small number of patients, and the results obtained may not be applicable to all incident hemodialysis patients. Specifically, the number of patients in some of the access-type-by-sex subgroups is limited (eg, 65 women with an AV fistula), and patients using a temporary access 1 month after starting chronic hemodialysis are excluded. The USRDS reports that 27% of incident hemodialysis patients are using a temporary access after 1 month of hemodialysis.39 In the population from which the study patients were drawn, 705 of the patients (37.0%) for whom the type of access was specified were using an indwelling catheter 1 month after starting hemodialysis therapy. Although a comparison of the characteristics of the study population to those of the remaining patients in the vascular access component of Wave I found a difference only in mean age, the study population may still represent a selected subset of healthier patients and/or those receiving better care before starting chronic hemodialysis therapy. Nevertheless, the incidence rate for first subsequent procedures in this population is very high (0.61 procedures/ access-year). Another limitation is that the outcome used in the present study (ie, first vascular access– related procedure after 1 month of hemodialysis therapy) does not represent the entire burden of vascular access complications in the hemodialysis population because multiple events in the same patient are not included. If differences exist in outcomes after these initial procedures by sex and/or access type, the conclusions regarding the rates of complications among men and women with AV fistulae and PTFE grafts could be significantly affected. This study has several other limitations that result from the design of the DMMS. Data were obtained through chart reviews conducted by dialysis facility personnel, but there were no standardized criteria for selecting reviewers
GENDER AND VASCULAR ACCESS EVENTS
within each facility based on clinical knowledge. Chart reviews are limited by the completeness of the medical record for the presence or absence of specific comorbidities, and the results were not validated. The DMMS does not include some potentially important covariates that may impact on intervention-free access survival, such as access location or prior placement of subclavian catheters. We were also unable to include measurements of vessel diameter or results of diagnostic procedures related to vascular access that may have influenced the type of access placed. The DMMS is also limited to those patients insured by Medicare and excludes patients with private or no insurance. However, the USRDS estimates that Medicare insures 93% of all patients with ESRD in the United States.40 In summary, this investigation of a large representative sample of incident hemodialysis patients shows a significant benefit from placing AV fistulae rather than PTFE grafts in men aged younger than 73 years who will undergo chronic hemodialysis. No benefit was observed in women or older men. These findings could be interpreted as supporting the need for sex- and age-specific recommendations regarding preferred access types. Furthermore, these data point out that AV fistula–associated complication rates are very high in women, suggesting great potential for improvement. A better understanding of the determinants of successful access maturation and maintenance during the first 3 months of hemodialysis therapy, particularly in women, is needed. REFERENCES 1. Feldman HI, Held PJ, Hutchinson JT, Stoiber E, Hartigan MF, Berlin JA: Hemodialysis vascular access morbidity in the United States. Kidney Int 43:1091-1096, 1993 2. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S160-S177, 1997 (suppl 1) 3. Goldwasser P, Avram MM, Collier JT, Michel MA, Gusik SA, Mittman N: Correlates of vascular access occlusion in hemodialysis. Am J Kidney Dis 24:785-794, 1994 4. Kinnaert P, Vereerstraeten P, Toussaint C, Van Geertruyden J: Nine years’ experience with internal arteriovenous fistulas for haemodialysis: A study of some factors influencing the results. Br J Surg 64:242-246, 1977 5. Mehta S: Statistical summary of clinical results of vascular access procedures for hemodialysis, in Sommer BG, Henry ML (eds): Vascular Access for Hemodialysis. Chicago, IL, Precept Press, 1991, pp 145-157 6. Munda R, First MR, Alexander JW, Linnemann CC Jr,
1133
Fidler JP, Kittur D: Polytetrafluoroethylene graft survival in hemodialysis. JAMA 249:219-222, 1983 7. Palder SB, Kirkman RL, Whittemore AD, Hakim RM, Lazarus JM, Tilney NL: Vascular access for hemodialysis. Patency rates and results of revision. Ann Surg 202:235-239, 1985 8. Zibari GB, Rohr MS, Landreneau MD, Bridges RM, DeVault GA, Petty FH, Costley KJ, Brown ST, McDonald JC: Complications from permanent hemodialysis vascular access. Surgery 104:681-686, 1988 9. Hirth RA, Turenne MN, Woods JD, Young EW, Port FK, Pauly MV, Held PJ: Predictors of type of vascular access in hemodialysis patients. JAMA 276:1303-1308, 1996 10. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S54-S66, 1997 (suppl 1) 11. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S67-S85, 1997 (suppl 1) 12. Windus DW, Jendrisak MD, Delmez JA: Prosthetic fistula survival and complications in hemodialysis patients: Effects of diabetes and age. Am J Kidney Dis 19:448-452, 1992 13. Feldman HI, Kobrin S, Wasserstein A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 7:523-535, 1996 14. Woods JD, Turenne MN, Strawderman RL, Young EW, Hirth RA, Port FK, Held PJ: Vascular access survival among incident hemodialysis patients in the United States. Am J Kidney Dis 30:50-57, 1997 15. Ifudu O, Mayers JD, Cohen LS, Paul H, Brezsnyak WF, Avram MM, Herman AI, Friedman EA: Correlates of vascular access and nonvascular access-related hospitalizations in hemodialysis patients. Am J Nephrol 16:118-123, 1996 16. Golledge J, Smith CJ, Emery J, Farrington K, Thompson HH: Outcome of primary radiocephalic fistula for haemodialysis. Br J Surg 86:211-216, 1999 17. Aman LC, Levin NW, Smith DW: Hemodialysis access site morbidity. Proc Clin Dial Transplant Forum 10:277-284, 1980 18. Churchill DN, Taylor DW, Cook RJ, LaPlante P, Barre P, Cartier P, Fay WP, Goldstein MB, Jindal K, Mandin H: Canadian Hemodialysis Morbidity Study. Am J Kidney Dis 19:214-234, 1992 19. Rocco MV, Bleyer AJ, Burkart JM: Utilization of inpatient and outpatient resources for the management of hemodialysis access complications. Am J Kidney Dis 28:250256, 1996 20. Prischl FC, Kirchgatterer A, Brandstatter E, Wallner M, Baldinger C, Roithinger FX, Kramar R: Parameters of prognostic relevance to the patency of vascular access in hemodialysis patients. J Am Soc Nephrol 6:1613-1618, 1995 21. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S21-S39, 1997 (suppl 1) 22. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Washington, DC, US Department of Health and Human Services, Public Health Service Health Care Financing Administration, 1991
1134
23. Physician’s Current Procedural Terminology 4. Chicago, IL, American Medical Association, 1992 24. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 25. Cox DR: Regression models and life tables. J R Stat Soc 34:187-220, 1972 26. SAS Institute: SAS User’s Guide: Statistics, version 5. Cary, NC, SAS Institute, 1985 27. StataCorp: Stata Statistical Software: Release 5.0. College Station, TX, StataCorp, 1997 28. Ifudu O, Macey LJ, Homel P, Hyppolite JC, Hong J, Sumrani N, Distant D, Sommer BG, Friedman EA: Determinants of type of initial hemodialysis vascular access. Am J Nephrol 17:425-427, 1997 29. Consensus Development Conference Panel: Morbidity and mortality of renal dialysis: An NIH consensus conference statement. Consensus Development Conference Panel. Ann Intern Med 121:62-70, 1994 30. National Kidney Foundation: DOQI Clinical Practice Guidelines for Vascular Access. Am J Kidney Dis 30:S150S191, 1997 (suppl 3) 31. Reilly DT, Wood RF, Bell PR: Prospective study of dialysis fistulas: Problem patients and their treatment. Br J Surg 69:549-553, 1982 32. Wong V, Ward R, Taylor J, Selvakumar S, How TV, Bakran A: Factors associated with early failure of arteriovenous fistulae for haemodialysis access. Eur J Vasc Endovasc Surg 12:207-213, 1996 33. Hussain KM, Kogan A, Estrada AQ, Kostandy G,
ASTOR ET AL
Foschi A, Dadkhah S: Referral pattern and outcome in men and women undergoing coronary artery bypass surgery—A critical review [published erratum appears in Angiology 49:870, 1998]. Angiology 49:243-250, 1998 34. Davis KB: Coronary artery bypass graft surgery in women, in Eaker AD, Packard B, Wenger NK (eds): Coronary Artery Disease in Women. New York, Haymarket Doyma, 1987, pp 247-250 35. Loop FD, Golding LR, MacMillan JP, Cosgrove DM, Lytle BW, Sheldon WC: Coronary artery surgery in women compared with men: Analyses of risks and long-term results. J Am Coll Cardiol 1:383-390, 1983 36. Eritsland J, Arnesen H, Fjeld NB, Gronseth K, Abdelnoor M: Risk factors for graft occlusion after coronary artery bypass grafting. Scand J Thorac Cardiovasc Surg 29:63-69, 1995 37. Windus DW: Permanent vascular access: A nephrologist’s view. Am J Kidney Dis 21:457-471, 1993 38. Wetzig GA, Gough IR, Furnival CM: One hundred cases of arteriovenous fistula for haemodialysis access: The effect of cigarette smoking on patency. Aust N Z J Surg 55:551-554, 1985 39. US Renal Data System: Excerpts From the US Renal Data System 1996 Annual Data Report. Am J Kidney Dis 28:S58-S78, 1996 (suppl 2) 40. US Renal Data System: Excerpts from US Renal Data System 1999 Annual Data Report. Am J Kidney Dis 34:S1-S3, 1999 (suppl 1)
V
ASCULAR ACCESS complications are the leading cause of morbidity in chronic hemodialysis patients, accounting for 15% of the hospitalizations in this population.1 Medicare From the Departments of Epidemiology, Biostatistics, and Health Policy and Management, The Johns Hopkins School of Hygiene and Public Health; Welch Center for Prevention, Epidemiology, and Clinical Research, The Johns Hopkins University; and the Department of Medicine, The Johns Hopkins University School of Medicine, Baltimore, MD. Received February 16, 2000; accepted in revised form June 30, 2000. Supported in part by grant no. T32HL07024-23 from the National Heart, Lung, and Blood Institute (B.C.A.); grants no. R29DK48362 (J.C.), K24DK02643 (N.R.P.), T32DK07732 (M.J.K.), and K24DK02856 (M.J.K.) from the National Institute of Diabetes and Digestive and Kidney Diseases; and grant no. R01-HS-08365 from the Agency for Health Care Research and Quality. The data reported here have been supplied by the US Renal Data System. The interpretation and reporting of these data are the responsibility of the author(s) and in no way should be seen as an official policy or interpretation of the US government. Address reprint requests to Josef Coresh, MD, PhD, 2024 E Monument St, Ste 2-600, Baltimore, MD 21205. E-mail: [email protected] © 2000 by the National Kidney Foundation, Inc. 0272-6386/00/3606-0005$3.00/0 doi:10.1053/ajkd.2000.19816 1126
expenditures for vascular access–related procedures were nearly $1 billion in 1994, representing between 14% and 17% of the total spending for hemodialysis patients.2 Most studies have found native arteriovenous (AV) fistulae to be associated with improved access survival and fewer complications than prosthetic polytetrafluoroethylene (PTFE) grafts.3-9 However, the proportion of patients starting chronic hemodialysis therapy in the United States who receive an AV fistula has declined in the last decade.9,10 In 1993, more than two thirds of the permanent accesses in use at 1 month after starting hemodialysis were PTFE grafts.11 The reasons for this trend are unclear but may be the increasing mean age at incidence of end-stage renal disease (ESRD) and the increasing proportion of patients with diabetes mellitus as the cause of ESRD.10,12 Age and diabetes mellitus, as well as female sex, have been associated with a greater likelihood of having a PTFE graft than an AV fistula in place 30 days after starting hemodialysis.9 Several factors have been postulated as predictive of vascular access complications, including having a PTFE graft rather than an AV fistula, black race, advanced age, diabetes mellitus, and a history of peripheral vascular disease.13,14
American Journal of Kidney Diseases, Vol 36, No 6 (December), 2000: pp 1126-1134
GENDER AND VASCULAR ACCESS EVENTS
Woods et al14 recently reported that the risk for an access complication requiring intervention increased significantly with increasing age for AV fistulae but not PTFE grafts. They found that AV fistulae and PTFE grafts had similar rates of complications in patients aged older than 65 years. Published data are conflicting on the effect of gender on rates of access-related complications. Several studies1,15,16 found greater rates of complications in women than men, whereas many others found no effect.8,14,17-19 Few of these studies examined the effect of gender among patients with AV fistulae separately from the effect among patients with PTFE grafts, and none has included a direct comparison of AV fistulae to PTFE grafts in men and women separately.16,17,20 The current study examines vascular access complications in a large cohort of incident hemodialysis patients who had a permanent access in use 1 month after starting hemodialysis therapy. The study compares complications in men and women and tests whether the benefit of AV fistulae over PTFE grafts is smaller in older individuals. In addition, the study examines the effects of clinical variables, including access type, on vascular access complications. METHODS
Study Design and Population We conducted a national longitudinal study of vascular access complications in incident hemodialysis patients. The study population was drawn from the 1,997 patients included in the vascular access component of Wave 1 of the Dialysis Mortality and Morbidity Study (DMMS) from the US Renal Data System (USRDS).21 The vascular access component of Wave 1 included all patients who started hemodialysis therapy during December 1993 at 550 randomly selected hemodialysis units and a random sample of patients from the same units who had started hemodialysis therapy earlier in 1993 and were undergoing hemodialysis on December 31, 1993. Our study included only those patients with a PTFE graft or AV fistula in use 1 month after starting hemodialysis therapy (n ⫽ 833). Patients were excluded for the following reasons: type of access in use was a temporary catheter (n ⫽ 705), bovine graft (n ⫽ 34), or missing (n ⫽ 91); the date of first chronic hemodialysis session was missing or was not during 1993 (n ⫽ 115); there was a lapse in all claims for greater than 60 days during the follow-up period (n ⫽ 131); or an access-related claim was reported before 1 month after starting hemodialysis therapy (n ⫽ 6). Because smoking was studied as a risk factor for subsequent vascular access procedures, an additional 82 patients were excluded for whom smoking information was missing.
1127
Data Collection Baseline data were obtained from the DMMS Standard Analysis File from the USRDS.21 Data were abstracted from patient medical records by dialysis facility personnel from March through July 1995. Data included information on patient demographics, comorbid conditions present in the last 10 years, psychosocial factors, laboratory values, dialysis treatments, type of vascular access in use 1 month after starting chronic hemodialysis therapy, and the date this access was placed. Race was categorized as black or other. Smoking was defined as any or no history of smoking. Additional analyses considered current smoking and previous smoking separately. Patients were considered to have a positive history of cardiovascular disease if they had undergone coronary angioplasty in the past or had been previously diagnosed with any of the following: angina, myocardial infarction, coronary heart disease, or coronary artery disease. Follow-up data for up to 4 years were obtained from detailed USRDS records that contained information on Medicare claims for hospital inpatient and outpatient services through December 31, 1996. Vascular access procedure codes from the International Classification of Diseases, Ninth Revision22 and the Physicians’ Current Procedural Terminology, Fourth Edition23 were included as outcomes. These codes correspond to creation of a new access, insertion of a venous catheter, and thrombectomy, revision, or removal of an access. The first date listed for a claim was used as the date of the access procedure.
Statistical Analysis Follow-up time was calculated starting 1 month after the initiation of chronic hemodialysis therapy until the date of the first access-related procedure recorded (hereafter referred to as the first subsequent procedure). Patients were censored at the time of death, renal transplantation, switch to peritoneal dialysis, or, if no event had occurred, on December 31, 1996 (the last date of follow-up). Time to the first subsequent procedure was analyzed using Kaplan-Meier estimates24 for the entire cohort and after stratification by access type, sex, and age. Cox proportional hazards models25 were used to examine the independent effect of sex on access complications after adjustment for demographic (age and race) and clinical (presence of diabetes mellitus or history of smoking, peripheral vascular disease, and cardiovascular disease) characteristics. The effect of age on the risk for a first subsequent procedure was further studied by dividing the male and female populations into four groups, each based on approximate quartiles of age. Access type– specific incidence rates and adjusted relative hazards associated with a PTFE graft versus an AV fistula were determined for each age group. Statistical analyses were performed using SAS (SAS Institute, Cary, NC)26 and Stata (StataCorp, College Station, TX).27
RESULTS
Study Population Characteristics The characteristics of patients who met the inclusion criteria for this study were compared
1128
ASTOR ET AL
with the 1,164 patients in Wave I who were excluded. The mean age of the study patients was 1.4 years older than the excluded patients (P ⬍ 0.005), but the groups did not differ in terms of sex, race, presence of diabetes mellitus, history of peripheral vascular disease, or history of cardiovascular disease. The study population consisted of 833 patients with a PTFE graft (n ⫽ 621) or AV fistula (n ⫽ 212) in use 1 month after starting hemodialysis therapy. Mean age at the time of first chronic hemodialysis session was 63.0 ⫾ 14.2 (SD) years. Nearly half the patients (406 patients) were women, and one third (281 patients) were black (Table 1). Patients receiving a PTFE graft were 4 years older (P ⬍ 0.001) and more likely to be women (P ⬍ 0.001), be black (P ⬍ 0.01), and have diabetes (P ⬍ 0.01). Patients with a PTFE graft were also more likely to have a history of peripheral vascular disease and cardiovascular disease, but these differences were not statistically significant (P ⫽ 0.14 and 0.06, respectively). Vascular Access Outcomes Follow-up through 1996 identified 552 first access–related procedures occurring more than 1 month after initiation of chronic hemodialysis therapy (ie, first subsequent procedures), consisting of the creation of 90 new accesses (16.3%), 59 venous catheter insertions (10.7%), and 403 Table 1. Characteristics by Access Type of 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis
Patient Characteristic
Age quartile 1 (18-54 y) 2 (55-64 y) 3 (65-72 y) 4 (ⱖ73 y) Women Black Diabetes mellitus Ever smoked History of peripheral vascular disease History of cardiovascular disease NOTE.
AV Fistula (n ⫽ 212)
PTFE Graft (n ⫽ 621)
69 (32.6) 42 (19.8) 54 (25.5) 47 (22.2) 65 (30.1) 56 (26.4) 90 (42.5) 99 (46.7) 44 (20.8)
139 (22.4) 141 (22.7) 174 (28.0) 167 (26.9) 341 (54.9) 225 (36.2) 330 (53.1) 277 (44.6) 160 (25.8)
77 (36.3)
272 (43.8)
P
⬍0.001 ⬍0.001 ⬍0.01 ⬍0.01 0.60 0.14
Values expressed as number (percent).
0.06
other procedures (73.0%; eg, thrombectomies, revisions, and removals). Overall median follow-up was 13.0 months. The incidence rate of first subsequent procedures was 0.61 procedures/ access-year. Of the 552 initial procedures identified, 442 occurred in PTFE grafts (incidence rate, 0.71 procedures/access-year) and 110 in AV fistulae (incidence rate, 0.39 procedures/accessyear). Univariate Associations With Initial Vascular Access Procedures AV fistulae had better procedure-free survival in men than women (P ⬍ 0.01; Fig 1). The median procedure-free survival among men with an AV fistula was 929 days compared with 329 days among women with an AV fistula. This sex difference was apparent (P ⬍ 0.02 by log rank) within the first 30 days of follow-up, by which time an access-related procedure had been performed in 14% of the women with an AV fistula compared with 5% of the men with an AV fistula. PTFE grafts had slightly better procedure-free survival in women than men (P ⬍ 0.05), with median procedure-free survival times of 332 and 267 days, respectively. Multivariate Associations With Initial Vascular Access Procedures Use of a PTFE graft compared with an AV fistula was associated with more than a twofold increase in the risk for an access-related procedure in multivariate proportional hazards analysis in men (relative hazard, 2.15; 95% confidence interval [CI], 1.62 to 2.85), but not in women (relative hazard, 1.03; 95% CI, 0.73 to 1.44), after adjustment for age, race, presence of diabetes, and history of smoking, peripheral vascular disease, and cardiovascular disease (Table 2). The association of peripheral vascular disease with initial access procedures reached statistical significance in men (relative hazard, 1.38; 95% CI, 1.09 to 1.67), but not in women (relative hazard, 1.19; 95% CI, 0.90 to 1.49). Conversely, the association of any history of smoking with initial access procedures reached statistical significance in women (relative hazard, 1.33; 95% CI, 1.07 to 1.59), but not in men (relative hazard, 1.18; 95% CI, 0.94 to 1.43). In a combined analysis of men and women, the presence of peripheral vascular disease (relative haz-
GENDER AND VASCULAR ACCESS EVENTS
1129
Fig 1. Kaplan-Meier estimates of freedom from first subsequent vascular access procedure by access type and sex for 833 incident hemodialysis patients with a permanent access in use 1 month after starting chronic hemodialysis therapy.
ard, 1.24; 95% CI, 1.01 to 1.52) and smoking (relative hazard, 1.25; 95% CI, 1.05 to 1.49) were associated with an increased risk for procedures after adjustment for the other factors in Table 2. Similar results were obtained from models considering current (n ⫽ 160) and previous (n ⫽ 216) smoking separately. Age, race, presence of diabetes mellitus, and history of cardiovascular disease did not predict first subsequent procedures in either sex or overall. In analyses
stratified by access type, female sex (versus male sex) was associated with a relative hazard of 1.58 (95% CI, 1.05 to 2.35) in patients with an AV fistula compared with a relative hazard of 0.90 (95% CI, 0.73 to 1.09) in patients with a PTFE graft. Modification of PTFE Graft Complication Risk by Age in Men and Women Among men, use of a PTFE graft compared with an AV fistula was associated in unadjusted
Table 2. Adjusted Relative Hazard of First Subsequent Vascular Access Procedure Associated With Patient Characteristics for 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis by Sex Men (n ⫽ 427)
P
Adjusted Relative Hazard
95% Confidence Interval
— 1.62 ⫺2.85 0.92 ⫺1.11 0.95 ⫺1.65 0.78 ⫺1.28 0.93 ⫺1.51 1.03 ⫺1.84
— ⬍0.001
1.0 1.03
0.85
1.01
0.11
1.09
0.99
0.99
0.18
1.33
0.03
1.19
— 0.73 ⫺1.44 0.92 ⫺1.11 0.86 ⫺1.39 0.76 ⫺1.28 1.03 ⫺1.72 0.89 ⫺1.60
0.63 ⫺1.09
0.17
1.21
Adjusted Relative Hazard
95% Confidence Interval
AV fistula PTFE graft
1.0 2.15
Age (per 10 y)
1.01
Black (v other race)
1.25
Presence of diabetes mellitus (v absence) Ever smoked (v never)
1.00
Patient Characteristic
Positive history of peripheral vascular disease (v negative history) Positive history of cardiovascular disease (v negative history)
1.18 1.38
0.83
Women (n ⫽ 406)
0.94 ⫺1.55
P
— 0.88 0.85 0.48 0.94 0.03 0.23
0.15
1130
ASTOR ET AL
analysis with a greater risk for first subsequent procedures in patients in the lower three quartiles of age (ie, ⬍73 years; P ⬍ 0.01 by log rank), but not in older patients (P ⫽ 0.67; Table 3). Among women, there was no statistically significant difference in risk among patients with PTFE grafts and those with AV fistulae in any age quartile. Multivariate proportional hazards analysis stratified by age quartile produced similar results after adjusting for race, presence of diabetes, smoking history, history of peripheral vascular disease, and history of cardiovascular disease. The relative hazard of first subsequent procedures associated with having a PTFE graft (compared with an AV fistula) varied with age quartile in men; 2.78 (95% CI, 1.66 to 4.65) in the lowest age quartile, 2.16 (95% CI, 1.10 to 4.26) in the second age quartile, 2.60 (95% CI, 1.26 to 4.46) in the third age quartile, and 1.02 (95% CI, 0.58 to 1.82) in the highest age quartile. Among women, the relative hazard was not significantly different from 1.0 in any group or overall and showed an inconsistent pattern across age groups. The number of AV fistulae in women (n ⫽ 65) limits the precision of these estimates. DISCUSSION
Previous studies have shown AV fistulae to have superior outcomes for hemodialysis patients compared with prosthetic grafts.3,5,8,18 The present study shows a significant benefit associ-
ated with use of an AV fistula in men, but not in women. These results held after adjusting for age, race, presence of diabetes mellitus, smoking history, history of peripheral vascular disease, and history of cardiovascular disease. This sexspecific benefit of AV fistulae over PTFE grafts has not been reported previously. An association of female sex with a greater risk for accessrelated hospitalization has been reported in two large studies,1,15 although neither investigated the results by access type. This distinction is important because it has been shown consistently that women are more likely to receive a PTFE graft than men.9,14,18,19,28 Many other studies found no association between sex and vascular access complications, but these also did not distinguish between AV fistulae and PTFE grafts.8,14,18,19 The effect of sex on access-related complications has also been examined in studies limited to patients using an AV fistula4,16,17,20 or to patients using a PTFE graft.17 For example, Aman et al17 found no sex differences in access failure or complication rates in 91 patients with PTFE grafts. Conversely, Golledge et al16 recently reported results from 107 patients (73 men, 34 women) with a first AV fistula; they found lower access survival for women after 2 years of follow-up (P ⬍ 0.01). Prischl et al20 found no significant differences in access survival between 80 men and 43 women on hemodialysis therapy with a first AV fistula
Table 3. Incidence Rate Per Access-Year of First Subsequent Vascular Access Procedure for 833 Incident Hemodialysis Patients With a Permanent Vascular Access in Use 1 Month After Starting Chronic Hemodialysis by Access Type, Sex, and Age Quartile Age Quartile
Men AV fistula (n ⫽ 147) PTFE graft (n ⫽ 280) P Women AV fistula (n ⫽ 65) PTFE graft (n ⫽ 341) P NOTE.
1 (18-54 y)
2 (55-64 y)
3 (65-72 y)
4 (ⱖ73 y)
Overall
0.29 (0.19-0.43)
0.28 (0.16-0.50)
0.29 (0.18-0.48)
0.58 (0.36-0.94)
0.33 (0.26-0.42)
0.99 (0.77-1.28)
0.72 (0.53-0.98)
0.83 (0.64-1.08)
0.65 (0.49-0.86)
0.79 (0.69-0.91)
⬍0.001
⬍0.001
⬍0.001
0.67
⬍0.001
0.41 (0.21-0.83)
0.52 (0.29-0.95)
0.60 (0.32-1.11)
0.96 (0.55-1.69)
0.59 (0.43-0.80)
0.70 (0.53-0.92)
0.55 (0.42-0.71)
0.78 (0.62-0.98)
0.60 (0.46-0.78)
0.65 (0.57-0.74)
0.21
0.34
0.14
0.14
0.74
Values expressed as incidence rate (95% CI). P by log rank.
GENDER AND VASCULAR ACCESS EVENTS
(P ⫽ 0.24), although the survival curves show poorer access survival for women at 1 year. Although all these studies have yielded valuable results, the evidence that may guide clinical practice is limited. Sample sizes have been small, and patient populations with AV fistulae were derived from single European centers. Differences in access placement practices between Europe and the United States limited the extent to which results could be generalized to the United States. Our study addresses these shortcomings, and the incidence rates for first access–related procedures occurring after 1 month of hemodialysis therapy (0.71 procedures/access-year for PTFE grafts; 0.39 procedures/access-year for AV fistulae) observed in this study are similar to those reported elsewhere.5,14,18,29,30 The results of previous studies are consistent with our findings that AV fistula–related procedure rates are greater in women than men, but PTFE graft–related procedure rates are similar by sex. However, because our study included both access types and had a larger sample, we were able to directly compare AV fistulae with PTFE grafts among men and women separately. Feldman et al1 suggested that differences in vessel diameter may account for the increased access-related morbidity experienced by women in their study. In a report of surgical placement of AV fistulae, Reilly et al31 found that vein size was a significant predictor of subsequent fistula survival, although sex was not. A more recent European study also found that small vessel size predicted fistula failure in the first 3 months after surgery.32 The effect of vessel size may be of less significance in patients with PTFE grafts than AV fistulae. This sex difference in vessel diameter has also been suggested as a reason for the poorer outcomes observed in women after coronary artery bypass surgery.33 In several studies of coronary artery bypass grafting that included measurements of vessel size, this characteristic was found to be the most predictive of outcome, diminishing any effect of sex.34-36 Sex differences in hemostatic factors have also been suggested as an explanation for the increased rate of vascular access complications in women.15 Our data show that women with AV fistulae had significantly more early access-related proce-
1131
dures (ie, during the first month of the study, which corresponds to the second month of hemodialysis therapy) than did their male counterparts. This result has been observed in previous studies4,16 and suggests that lack of successful access maturation is a critical element in the greater rate of complications associated with AV fistulae in women. Our study confirms the findings of Woods et al14 in the USRDS Case Mix Adequacy Study population in that the benefit of AV fistulae over PTFE grafts was found only in younger patients. However, the benefit we observed was exclusive to men. Previous studies have investigated the association of age with access complications but did not consider the effects of access type or sex. Several previous reports failed to find an association between age and access complications,8,18,19 although others reported a significant effect of age appearing after 6 months of follow-up20 and in patients without diabetes with prosthetic grafts.12 A review of USRDS data reported a significantly greater rate of access-related hospitalizations among those aged older than 64 years,1 although access type was not available for these patients. Woods et al14 reported a relative hazard of 1.40 for access procedures associated with a 10-year older age for an AV fistula in patients aged younger than 65 years and no significant effect of age in older patients or in those with a PTFE graft. The adjusted relative hazard of a PTFE graft versus an AV fistula decreased in that study from 3.0 at the age of 40 years to 1.3 at the age of 65 years and older. Among men in the present study, use of a PTFE graft compared with an AV fistula was associated with significantly lower procedure-free access survival in the three younger age quartiles; however, in the oldest age quartile (ⱖ73 years), the benefit was smaller and not statistically significant. The present study also confirms an earlier report of a significant association between a history of peripheral vascular disease and subsequent access procedures.14 This association is not surprising because the condition of the native vasculature is likely a determinant of the adequacy of any access created. The pathophysiological characteristics of peripheral vascular disease and vascular access complications may also be similar. It has been postulated that periph-
1132
eral vascular disease may compromise blood flow in the access, producing a greater risk for thrombosis.37 Smoking history has been examined in a few studies of vascular access morbidity, with inconsistent results.18,31,38 In this study, current or previous smoking was associated with a relative hazard of 1.25 (95% CI, 1.05 to 1.49) for subsequent access events. Analyzed separately, the relative hazard of previous smoking was nearly identical to that of current smoking. Previous peripheral vascular damage in former and active smokers, leading to acute access thrombosis, may partially explain this observation. The results of the present study have several important implications. The finding that AV fistulae have no benefit in terms of procedure-free access survival in women and older men suggests that the preferred type of permanent vascular access for all patients may not be as straightforward as described in the National Kidney Foundation’s Dialysis Outcomes Quality Initiative clinical practice guidelines.30 Whether any long-term benefits, such as improved access survival after interventions, are gained by women and older men with AV fistulae needs to be investigated in prospective studies. Such studies ideally should include data on the creation and maintenance of permanent accesses before the initiation of chronic hemodialysis therapy to evaluate the entire benefit of creating a functioning AV fistula in these patients. The results of these studies may show that a different costbenefit analysis regarding the optimal choice of access type is necessary for patients with different characteristics. The results of the present study also have implications relating to screening for access dysfunction. Compared with men with an AV fistula, women with an AV fistula experienced significantly more interventions during the first 3 months of hemodialysis therapy in our study. This finding suggests that women may require more frequent screening for early fistula dysfunction and that this frequent screening should continue for at least 3 months after the initiation of hemodialysis therapy. This finding may also suggest that women require a longer time for fistula maturation before its first use. During the present study (1993 to 1996), it is probable that few of
ASTOR ET AL
the participating centers had consistent access screening programs. Thus, it is unlikely that the greater rates of access-related procedures observed in women were caused by access problems detected during routine screening. This study has several limitations. Despite the large size and representative nature of the Wave 1 patient population, some subgroups analyzed in this study include a relatively small number of patients, and the results obtained may not be applicable to all incident hemodialysis patients. Specifically, the number of patients in some of the access-type-by-sex subgroups is limited (eg, 65 women with an AV fistula), and patients using a temporary access 1 month after starting chronic hemodialysis are excluded. The USRDS reports that 27% of incident hemodialysis patients are using a temporary access after 1 month of hemodialysis.39 In the population from which the study patients were drawn, 705 of the patients (37.0%) for whom the type of access was specified were using an indwelling catheter 1 month after starting hemodialysis therapy. Although a comparison of the characteristics of the study population to those of the remaining patients in the vascular access component of Wave I found a difference only in mean age, the study population may still represent a selected subset of healthier patients and/or those receiving better care before starting chronic hemodialysis therapy. Nevertheless, the incidence rate for first subsequent procedures in this population is very high (0.61 procedures/ access-year). Another limitation is that the outcome used in the present study (ie, first vascular access– related procedure after 1 month of hemodialysis therapy) does not represent the entire burden of vascular access complications in the hemodialysis population because multiple events in the same patient are not included. If differences exist in outcomes after these initial procedures by sex and/or access type, the conclusions regarding the rates of complications among men and women with AV fistulae and PTFE grafts could be significantly affected. This study has several other limitations that result from the design of the DMMS. Data were obtained through chart reviews conducted by dialysis facility personnel, but there were no standardized criteria for selecting reviewers
GENDER AND VASCULAR ACCESS EVENTS
within each facility based on clinical knowledge. Chart reviews are limited by the completeness of the medical record for the presence or absence of specific comorbidities, and the results were not validated. The DMMS does not include some potentially important covariates that may impact on intervention-free access survival, such as access location or prior placement of subclavian catheters. We were also unable to include measurements of vessel diameter or results of diagnostic procedures related to vascular access that may have influenced the type of access placed. The DMMS is also limited to those patients insured by Medicare and excludes patients with private or no insurance. However, the USRDS estimates that Medicare insures 93% of all patients with ESRD in the United States.40 In summary, this investigation of a large representative sample of incident hemodialysis patients shows a significant benefit from placing AV fistulae rather than PTFE grafts in men aged younger than 73 years who will undergo chronic hemodialysis. No benefit was observed in women or older men. These findings could be interpreted as supporting the need for sex- and age-specific recommendations regarding preferred access types. Furthermore, these data point out that AV fistula–associated complication rates are very high in women, suggesting great potential for improvement. A better understanding of the determinants of successful access maturation and maintenance during the first 3 months of hemodialysis therapy, particularly in women, is needed. REFERENCES 1. Feldman HI, Held PJ, Hutchinson JT, Stoiber E, Hartigan MF, Berlin JA: Hemodialysis vascular access morbidity in the United States. Kidney Int 43:1091-1096, 1993 2. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S160-S177, 1997 (suppl 1) 3. Goldwasser P, Avram MM, Collier JT, Michel MA, Gusik SA, Mittman N: Correlates of vascular access occlusion in hemodialysis. Am J Kidney Dis 24:785-794, 1994 4. Kinnaert P, Vereerstraeten P, Toussaint C, Van Geertruyden J: Nine years’ experience with internal arteriovenous fistulas for haemodialysis: A study of some factors influencing the results. Br J Surg 64:242-246, 1977 5. Mehta S: Statistical summary of clinical results of vascular access procedures for hemodialysis, in Sommer BG, Henry ML (eds): Vascular Access for Hemodialysis. Chicago, IL, Precept Press, 1991, pp 145-157 6. Munda R, First MR, Alexander JW, Linnemann CC Jr,
1133
Fidler JP, Kittur D: Polytetrafluoroethylene graft survival in hemodialysis. JAMA 249:219-222, 1983 7. Palder SB, Kirkman RL, Whittemore AD, Hakim RM, Lazarus JM, Tilney NL: Vascular access for hemodialysis. Patency rates and results of revision. Ann Surg 202:235-239, 1985 8. Zibari GB, Rohr MS, Landreneau MD, Bridges RM, DeVault GA, Petty FH, Costley KJ, Brown ST, McDonald JC: Complications from permanent hemodialysis vascular access. Surgery 104:681-686, 1988 9. Hirth RA, Turenne MN, Woods JD, Young EW, Port FK, Pauly MV, Held PJ: Predictors of type of vascular access in hemodialysis patients. JAMA 276:1303-1308, 1996 10. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S54-S66, 1997 (suppl 1) 11. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S67-S85, 1997 (suppl 1) 12. Windus DW, Jendrisak MD, Delmez JA: Prosthetic fistula survival and complications in hemodialysis patients: Effects of diabetes and age. Am J Kidney Dis 19:448-452, 1992 13. Feldman HI, Kobrin S, Wasserstein A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 7:523-535, 1996 14. Woods JD, Turenne MN, Strawderman RL, Young EW, Hirth RA, Port FK, Held PJ: Vascular access survival among incident hemodialysis patients in the United States. Am J Kidney Dis 30:50-57, 1997 15. Ifudu O, Mayers JD, Cohen LS, Paul H, Brezsnyak WF, Avram MM, Herman AI, Friedman EA: Correlates of vascular access and nonvascular access-related hospitalizations in hemodialysis patients. Am J Nephrol 16:118-123, 1996 16. Golledge J, Smith CJ, Emery J, Farrington K, Thompson HH: Outcome of primary radiocephalic fistula for haemodialysis. Br J Surg 86:211-216, 1999 17. Aman LC, Levin NW, Smith DW: Hemodialysis access site morbidity. Proc Clin Dial Transplant Forum 10:277-284, 1980 18. Churchill DN, Taylor DW, Cook RJ, LaPlante P, Barre P, Cartier P, Fay WP, Goldstein MB, Jindal K, Mandin H: Canadian Hemodialysis Morbidity Study. Am J Kidney Dis 19:214-234, 1992 19. Rocco MV, Bleyer AJ, Burkart JM: Utilization of inpatient and outpatient resources for the management of hemodialysis access complications. Am J Kidney Dis 28:250256, 1996 20. Prischl FC, Kirchgatterer A, Brandstatter E, Wallner M, Baldinger C, Roithinger FX, Kramar R: Parameters of prognostic relevance to the patency of vascular access in hemodialysis patients. J Am Soc Nephrol 6:1613-1618, 1995 21. US Renal Data System: Excerpts From the US Renal Data System 1997 Annual Data Report. Am J Kidney Dis 30:S21-S39, 1997 (suppl 1) 22. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Washington, DC, US Department of Health and Human Services, Public Health Service Health Care Financing Administration, 1991
1134
23. Physician’s Current Procedural Terminology 4. Chicago, IL, American Medical Association, 1992 24. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 25. Cox DR: Regression models and life tables. J R Stat Soc 34:187-220, 1972 26. SAS Institute: SAS User’s Guide: Statistics, version 5. Cary, NC, SAS Institute, 1985 27. StataCorp: Stata Statistical Software: Release 5.0. College Station, TX, StataCorp, 1997 28. Ifudu O, Macey LJ, Homel P, Hyppolite JC, Hong J, Sumrani N, Distant D, Sommer BG, Friedman EA: Determinants of type of initial hemodialysis vascular access. Am J Nephrol 17:425-427, 1997 29. Consensus Development Conference Panel: Morbidity and mortality of renal dialysis: An NIH consensus conference statement. Consensus Development Conference Panel. Ann Intern Med 121:62-70, 1994 30. National Kidney Foundation: DOQI Clinical Practice Guidelines for Vascular Access. Am J Kidney Dis 30:S150S191, 1997 (suppl 3) 31. Reilly DT, Wood RF, Bell PR: Prospective study of dialysis fistulas: Problem patients and their treatment. Br J Surg 69:549-553, 1982 32. Wong V, Ward R, Taylor J, Selvakumar S, How TV, Bakran A: Factors associated with early failure of arteriovenous fistulae for haemodialysis access. Eur J Vasc Endovasc Surg 12:207-213, 1996 33. Hussain KM, Kogan A, Estrada AQ, Kostandy G,
ASTOR ET AL
Foschi A, Dadkhah S: Referral pattern and outcome in men and women undergoing coronary artery bypass surgery—A critical review [published erratum appears in Angiology 49:870, 1998]. Angiology 49:243-250, 1998 34. Davis KB: Coronary artery bypass graft surgery in women, in Eaker AD, Packard B, Wenger NK (eds): Coronary Artery Disease in Women. New York, Haymarket Doyma, 1987, pp 247-250 35. Loop FD, Golding LR, MacMillan JP, Cosgrove DM, Lytle BW, Sheldon WC: Coronary artery surgery in women compared with men: Analyses of risks and long-term results. J Am Coll Cardiol 1:383-390, 1983 36. Eritsland J, Arnesen H, Fjeld NB, Gronseth K, Abdelnoor M: Risk factors for graft occlusion after coronary artery bypass grafting. Scand J Thorac Cardiovasc Surg 29:63-69, 1995 37. Windus DW: Permanent vascular access: A nephrologist’s view. Am J Kidney Dis 21:457-471, 1993 38. Wetzig GA, Gough IR, Furnival CM: One hundred cases of arteriovenous fistula for haemodialysis access: The effect of cigarette smoking on patency. Aust N Z J Surg 55:551-554, 1985 39. US Renal Data System: Excerpts From the US Renal Data System 1996 Annual Data Report. Am J Kidney Dis 28:S58-S78, 1996 (suppl 2) 40. US Renal Data System: Excerpts from US Renal Data System 1999 Annual Data Report. Am J Kidney Dis 34:S1-S3, 1999 (suppl 1)