It was recently reported that longer treatment reduced mortality in the DOPPS database for three large geographic areas: US, Europe and Japan [3]. The authors concluded that the data supported the generalized conclusion that mortality rate fell by 7% for each 30-min increase in treatment time. The data reported for each region adjusted for comorbidities are plotted in figure 12A and the statistical conclusion of a 7% decrease in mortality for each 30-min increase in treatment time is shown in figure 12B. It is not reasonable to generalize these data with such striking geographic differences in the effect of treatment time. The effect is almost negligible in the US and quite striking in Japan while mortality is much lower in Japan at all levels of treatment time.

Twardowski [9] has argued that 'short thrice weekly hemodialysis is inadequate regardless of small molecule clearance' and recommends 5- to 8-hour treatment time. Evaluation of this belief with the DOPPS data requires

• US A Europe □ Japan Statistical melting pot

Negative mortality <-►

.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Dialysis time (h)

US A Europe □ Japan -Statistical melting pot

Domain of immortal treatment times?

Negative mortality

Negative mortality

.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 Dialysis time (h)

Polynomial A Categorical — Linear (categorical)

Polynomial A Categorical — Linear (categorical)

Fig. 14. Results of Australia New Zealand observational study of outcome as function of treatment time and Kt/V. The categorical analyses show the typical tell-tale signs of dose targeting error for both Kt/V and treatment time. The polynomial fit shows U-shaped outcome responses for both parameters. Which is correct and why should there be a U shape?

Fig. 13. Extrapolation of DOPPS data to a treatment time of 8 h. A Domain of mortality. B Domain according to Twardowski [9].

Fig. 14. Results of Australia New Zealand observational study of outcome as function of treatment time and Kt/V. The categorical analyses show the typical tell-tale signs of dose targeting error for both Kt/V and treatment time. The polynomial fit shows U-shaped outcome responses for both parameters. Which is correct and why should there be a U shape?

expansion of the X axis to cover the disputed treatment time range from 2.0 to 8.0 h as shown in figure 13A. Clearly the DOPPS data do not cover such a wide range and, in fact, the mortality rate becomes negative when the DOPPS regression is extended to a treatment time of >5.5 h as shown in figure 13B. Twardowski's [9] belief that 8-hour treatment times are required for adequate clinical outcome might actually be expected to achieve immortality when viewed from a geometric interpretation of DOPPS data outcome.

It was also concluded in the DOPPS paper that slower ultrafiltration rates correlated with lower mortality. This is difficult to understand when in DOPPS II the mean ultrafiltration rare was 9.9 in Japan and 9.8 in the US [3] (table 1).

Was this article helpful?

## Post a comment