Modality switch

2.9.1 Epidemiology of modality switch

After having started the treatment, switching of modality is not infrequent.

Especially patients on PD commonly experience treatment failures resulting in modality change. In a register based study, less than 80% of 1390 dialysis patients remained on their first type of dialysis after two years (Smith and Wheeler 1987).

Gentil et al reported three-year technique survival (remaining on modality) 94%

and 56% on HD and PD, respectively (Gentil et al. 1991). An Italian study found method survival 40% in PD and over 90% in HD, follow-up continued up to nine years (Maiorca et al. 1996) whereas a study from the United States reported 30%

treatment failure on PD over 12 months follow-up (Singh et al. 1992). Jaar et al.

found that 25% of patients receiving PD and 5% of those on HD switched treatment modality at least once within seven years (Jaar et al. 2005). Shih et al.

conducted a register-based survey of over 3400 dialysis patients. Approximately one third of patients who started on PD had one or more modality changes within the first three years compared with less than 1% of the HD group (Shih et al.

2005). In a North American cohort study including 1587 PD patients, 39% had been forced to switch to HD by three years for at least 30 days (Shen et al. 2013).

Changes from one modality to another occur and multiple switches over the course of time are not uncommon. Intention-to-treat and as-treated approaches are applied when patient’s modality is classified. When the assignment of modality is done on an intention-to-treat basis, a modality is defined at study start and assignment does not change even if the patient switches to another modality. By intention-to-treat approach, bias resulting from high costs associated with modality switch is avoided. For example, patients who are performed an unsuccessful kidney transplantation and continue dialysis therapy, may have very high costs caused by failure-associated events. If these patients are not regarded as transplant patients and only those with functioning grafts are included in the analysis, costs of transplantation would be underestimated. Contrary, an as-treated approach – defined as modality actually in use at any point in time – would classify these patients according to new modality rather than to the failed modality that brought about the high costs.

2.9.2 Reasons for modality switch

Switching between modalities may occur for a variety of reasons. Potential causes of transfer from PD to HD can be classified as modality related (infections, inadequate dialysis and catheter-related problems), system related (lack of infrastructure or patient education, reimbursement policy) and patient related (comorbid conditions, burnout, social reasons, hernia formation and abdominal surgery) (Chaudhary et al. 2011). Of all causes, modality-related complications are the most common.

The development of PD-devices and supplies has helped to decrease the access-related infection rates (Strippoli et al. 2004) but peritonitis is still the main reason to change from PD to HD, especially within the first two years after initiation of treatment. In Canadian data from 1981 to 1997, the switch rate from PD was estimated to be 154 per 1000 patient-years and trend of a decreasing rate over years were found (Schaubel et al. 2001). A multicenter prospective study evaluated 292 PD patients. Almost 25% of patients experienced a treatment failure, 70% of them within two years of starting PD. Infections (peritonitis and catheter related infections) were the reason for the switch at 37% (Jaar et al. 2009). Peritonitis rates have varied markedly in published studies. In recent literature, an incidence rate ranging between one episode per 12–83 months have been reported (Monteon et al. 1998; Daly et al. 2001; Davenport 2009; Brown et al. 2011; Cnossen et al. 2011;

Hsieh et al. 2013; Ortiz et al. 2004; Medani et al. 2012). Insufficient dialysis or management of fluids is not infrequent. The ultrafiltration failure rates from 2% to 14% have been recognized (Jager et al. 1999). Mechanical complications may occur and contribute to switch from PD. Sometimes a patient transfers voluntarily to HD. Overall; trends toward better technique survival in PD over years have been reported (Guo and Mujais 2003).

Reasons for switching from HD to PD include vascular access problems and cardiovascular instability and also patient's own choice. Among HHD patients, lack of confidence in carrying out treatment, interference with home life, family dynamics and fear of self-cannulation may result in modality switch. In the UK, 166 HHD patients were followed for on average 2,3 years. Technique survivals at 1, 2 and 5 years were 98,4, 95,4 and 88,9%. Diabetes and cardiac failure associated with an increased risk of technique failure, but majority of patients switched modality for non-medical reasons (Jayanti et al. 2013).

A cohort study from the United States investigated the determinants of PD technique failure. Over 1500 PD patients were followed for three years. In multivariate analysis, female sex was statistically significantly associated with lower risk of failure whereas risk was increased in blacks, retired and disabled.

Interestingly, age, comorbidities, body mass index and educational or marital status did not associate with failure risk (Shen et al. 2013). Contrary, a recent study from Taiwan reported age over 65 to be the only identified risk factor for peritonitis, which subsequently predicted technique failure (Hsieh et al. 2013). In several studies, number of patients in PD centre has positively correlated with better outcomes (Schaubel et al. 2001; Huisman et al. 2002; Afolalu et al. 2009; Plantinga et al. 2009). High number of PD patients in each centre possibly accounts for the

excellent two-year technique survival rate of 82% in Hong Kong (Li and Szeto 2008).

2.9.3 Modality switch and costs

Substantial cost increase may be caused by switches between modalities. Initiation of a new modality is associated with start-up costs which accumulate in repeated switches (Prichard 1997). In a register-based study evaluating Medicare expenditures in dialysis patients over three years, annual costs were significantly lower (23% to 27%) for those patients who were treated on PD as the initial modality. Patients who switched from PD to HD within the first year had higher treatment costs than those who switched later and the economic advantage related to PD was lost among first-year switchers. In those patients who switched to HD for at least 60 days, annual costs increased over (US$) 20 000 (Shih et al. 2005). In a Canadian study, incident dialysis patients were categorized by initial modality and subsequent modality changes. Total treatment costs during three years were evaluated and a purchaser's perspective was taken. Adjusted cumulative three-year costs (Canadian $) were 58 724 for patients who received only PD and 175 996 for those who received only HD. Compared with patients on HD-only, costs were similar for patients with a PD technique failure and a change to HD. Costs for patients who changed from HD to PD were in between costs for PD-only and HD-only. The authors concluded that since costs were lower on PD and costs of patients with PD technique failure were not in excess of HD-only patients, the economic rationale for a PD-first policy in eligible patients was supported (Chui et al. 2012).

In HHD, a patient's inability to carry out dialysis procedure leads to technique failure and necessitates a modality change. In a Canadian study utilizing Markov model, annual probability of technique failure in nocturnal HHD was set at 7,6%

and HHD dominated CHD. HHD led to incremental cost savings (Canadian

$6700) and an additional 0,38 QALYs. In sensitivity analyses, higher risk of failure markedly increased costs. At an annual failure risk of 19%, CER 75 000/QALY was provided and its attractiveness was lost (Klarenbach et al. 2013). The reported failure rates have been lower in observational studies. In the UK, during an 8-year follow-up including 4528 HHD patient-months, technique failures were uncommon. Technique survivals at 1 and 5 years were 98,4% and 88,9%,

respectively. Age over 60 years, cardiac failure, diabetes and high score in comorbidity index scale were associated with technique failure (Jayanti et al. 2013).