The risk of hospital admission in home hemodialysis patients was similar to the risk in comparable in-center hemodialysis patients in a recent study published in the American Journal of Kidney Diseases. However, detailed analysis of hospital claims suggested that the similarity had an important caveat: HHD patients were hospitalized less frequently than IHD counterparts for cardiovascular complications, whereas HHD patients were hospitalized more frequently for infectious complications.

HHD has grown steadily in the US during the past decade. According to the ESRD Networks’ 2013 Annual Reports, there were 7,811 HHD patients in the US at the end of 2013, an increase of 8.4% from just the previous year. “The size of this population is now large enough to permit well-designed, observational studies that assess the benefits and risks of HHD across the entire country, rather than in isolated HHD programs,” says Eric Weinhandl, MS, senior epidemiologist at the Chronic Disease Research Group (CDRG) in Minneapolis, MN, and lead author of the study. “Our study of HHD patients is the largest yet.”

In “Hospitalization in daily home hemodialysis and matched thrice-weekly in-center hemodialysis patients,” Weinhandl and colleagues examined the risk of hospital admission in a cohort of 3,480 patients that initiated HHD between 2006 and 2009. All HHD patients were Medicare beneficiaries that dialyzed 5 to 6 times weekly with the NxStage System One. For each HHD patient, the authors identified 5 matched in-center hemodialysis patients in the United States Renal Data System (USRDS) registry. IHD patients were identified on the same day as the date of HHD initiation in addition to comparability in 33 factors, including age, race, sex, comorbid conditions, and dialysis provider. “Our goal was to identify 5 IHD patients that appeared to be HHD candidates and might have started HHD on the same day that another patient actually did, but instead received hemodialysis in a facility,” Weinhandl says.

In intention-to-treat analysis, almost 63% of HHD patients had been hospitalized after one year of follow-up, whereas 58% of IHD patients had been hospitalized. Overall, the relative hazard of hospital admission for HHD versus IHD was 1.01 (95% CI, 0.98-1.03). In on-treatment analysis, the relative hazard was similar. “The top-line finding is clearly neutral,” Weinhandl notes.

A closer look at cause of hospitalizations

When hospital admissions were categorized by the principal diagnosis at discharge, the study was much more illustrative. For admissions due to cardiovascular complications, the relative hazard of hospitalization for HHD versus IHD was 0.89 (0.86-0.93). The largest differences in risk were observed for admissions due to heart failure (hazard ratio, 0.69), cerebrovascular disease (0.85), and hypertensive disease (0.88), which collectively accounted for about half of admissions due to cardiovascular complications among matched IHD patients.

Allan Collins, MD, FACP, director of the Chronic Disease Research Group and senior author of the study, said, “The findings regarding cardiovascular risk on HHD are consistent with results from the Frequent Hemodialysis Network trial, which showed improved blood pressure control and reduced left ventricular mass for 6 versus 3 dialysis sessions weekly. Volume control is a core issue for dialysis patients, as it contributes to persistent hypertension and eventual heart failure. More frequent dialysis in indicated patients should be given wider consideration.”

In contrast, for admissions due to infectious complications, the relative hazard of hospitalization for HHD versus IHD was 1.18 (1.13-1.23). The largest differences in risk were observed for cardiac infection (hazard ratio, 3.42), osteomyelitis (1.48), vascular access infection (1.39), and bacteremia/sepsis (1.35). Despite the observational nature of the study, Weinhandl thinks that the data are unlikely to be spurious. “Differences in vascular access modality might confound our comparisons, but there are no national data that suggest that use of catheters is more common among HHD patients than among IHD patients.” Collins agreed, saying, “Infectious complications have been a perennial problem for HHD programs since the 1970s.”

Tracing the link to infection

The authors discussed several mechanisms that may underlie increased risk of infection on HHD, including dialysis frequency, buttonhole cannulation, and infection control practice in the home. Weinhandl suggests that more frequent dialysis might have an indirect effect on infection risk. “Plausibly, patients and care partners could become fatigued by the dialysis schedule, resulting in gradual deterioration of infection control practices. However, evidence of a physiologic link between dialysis frequency and infection risk is sparse and the quality of that evidence is low.”

Buttonhole cannulation may also contribute to increased risk in HHD patients with a fistula. The authors cited a systematic review that found more than 3 times the risk of access infection with buttonhole versus rope-ladder cannulation. Weinhandl notes that another review by Ben Wong, MD, and colleagues, published online in the American Journal of Kidney Diseases in August, found no significant difference in cannulation pain among randomized trials of buttonhole versus rope-ladder cannulation. “Many HHD programs and patients have successfully used buttonhole method. Thus, each dialysis provider should critically appraise whether buttonhole cannulation, as it is presently used, is a net benefit or a net harm to its HHD patients.”

Perhaps the most important factor is technique in the home environment. Good technique starts with good training. “HHD training programs are heavily focused on technical aspects, including setup and operation of the dialysis machine. Infection control practice may not be adequately addressed before HHD begins in the home,” Collins said. To this point, the study authors cited a presentation that Leslie Spry, MD, FACP, delivered at the past Annual Dialysis Conference. Spry and colleagues surveyed 55 HHD training clinicians and 129 HHD patients for adherence to “generally accepted practices” (GAP) regarding vascular access care. The majority of clinicians reported not teaching GAP regarding skin antisepsis. Even when training is adequate, infection control practices in the home may not be followed. The vast majority of patients in the survey reported not adhering to GAP regarding skin antisepsis, roughly 50% of patients with a catheter reported not adhering to GAP regarding exit site care, and about 25% of all patients indicated a lack of recognition and reporting of signs and symptoms of infection. “Infection control needs to be addressed at every monthly visit with the patient. Prevention is the most important approach,” said Collins. “Every HHD program should have an infection monitoring system in place, too.”

Being proactive about identifying infection

Collins suggests that there are several opportunities for nephrologists to gauge the presence of infection linked to the vascular access. “Temperature rise during the hemodialysis treatment was first described by Stanley Shaldon more than 25 years ago, yet very little has been done to assess its utility in infection control. Temperature rise may be a sign of subclinical infection.” Hepcidin may be another useful marker to consider. “A rise in hepcidin is a physiologic response to acute infection. Thus, outlying values in a time course of hepcidin measurements may be informative.” Collins argues that both of these methods should be rigorously evaluated for their diagnostic utility, not only in HHD, but also in IHD and peritoneal dialysis (PD), on which rates of hospital admission due to infection have been stubbornly high for many years.

Making changes to reduce vascular access-related infections

The authors are optimistic that the risk of infection on HHD is modifiable. They point to PD as a case in point: the rate of peritonitis fell sharply from 1980 to 1995 with accumulated experience and continued innovation. “Although HHD is not new per se, HHD programs in many end-stage renal disease facilities are young,” the authors wrote. Weinhandl notes that he and his colleagues found that while the risk of first admission due to infection was higher for HHD versus IHD, the risk of subsequent admission due to infection was similar for the modalities. “One possibility is that providers intervened after the first major infection on HHD,” said Weinhandl.

With persistent growth in HHD, the clinical significance of associations with both cardiovascular and infectious complications is obvious, as cardiovascular disease and infection are the leading causes of death in dialysis patients, adds Weinhandl. “The success of HHD likely depends on both patients and providers recognizing the advantages and inherent challenges of the modality.”