To the Editor:
Bone marrow transplantation (BMT) has been shown to effectively reduce the cognitive deterioration normally associated with Hurler syndrome.1 In their recent review of the Storage Disease Collaborative Study Group's retrospective experience using unrelated donor BMT for children with Hurler syndrome, Peters et al2 suggest that changes in post-BMT neuropsychological status differ as a function of pretransplant developmental level. Based on a sample of 11 children for whom both pre-BMT baseline and post-BMT neuropsychological testing was available, Peters et al2 conclude that children whose pre-BMT mental developmental index (MDI) is less than 70 show deterioration in their developmental skills, whereas children with a baseline MDI of greater than 70 show a “favorable long-term outcome and improved cognitive function” (abstract). We feel that this conclusion is premature given the limited size of the present sample and, furthermore, because the trends observed in their published data do not appear to support it.
Using the data presented in Table 3 Peters et al,2 we calculated that the 6 patients in the greater than 70 group showed a mean pre-BMT MDI of 91 and a mean post-BMT MDI of 75, indicating a mean decline of 16 points or roughly one standard deviation on the Bayley scale. The same calculation cannot be made for the less than 70 group because the test instrument does not differentiate MDIs less than 50. However, one can calculate a developmental quotient (DQ) for both groups by, taking the age equivalent scores, dividing by the patients chronological age at the time of testing, and then multiplying the resulting proportion by 100. Using this method, we calculated that the greater than 70 group shows a mean pre-BMT DQ of 90 and a post-BMT DQ of 71, a mean decline of 19 points. The less than 70 group shows a mean pre-BMT DQ of 46 and post-BMT DQ of 29, a mean decline of 17 points. Analyzed in this manner, the data indicate an equivalent decline in developmental function in both groups. Given the small numbers involved, we also analyzed the data in Table 3 on a case by case basis. We categorized developmental outcomes as stable (ie, any positive changes or declines of <15 points in DQ), minor deterioration (declines of 15 to 25 points in DQ), and major deterioration (declines of >25 points in DQ). Of the 6 patients in the greater than 70 group, 2 showed stable outcomes (PI nos. 1 and 30), 2 showed minor deterioration (PI nos. 22 and 32), and 2 showed major deterioration (PI nos. 4 and 15). Similarly, of the 5 patients in the less than 70 group, 1 showed stable function (PI no. 23), 2 showed minor deterioration (PI nos. 5 and 19), and 2 showed major deterioration (PI nos. 8 and 27).
We feel that the data from the report of Peters et al2 are not sufficient to allow firm conclusions, but suggest that children transplanted for Hurler syndrome experience a mild to moderate developmental decline after BMT regardless of premorbid developmental level. This decline in developmental function is not as severe as that seen in children with Hurler syndrome who did not receive transplants; thus, the benefits of BMT are indicated.1 The less than 70 group show a much poorer ultimate outcome, but this is due primarily to their significantly lower function pre-BMT and not to a difference in post-BMT developmental trajectory. The declines in the less than 70 group may appear more dramatic, particularly when children who are mildly to moderately developmentally delayed initially become severely delayed, and one could argue that the decline in quality of life is more pronounced for this group. Clearly, the major message from this study is that, for optimal developmental outcome, children with Hurler syndrome should receive transplants as young as possible, preferably before the onset of any cognitive deterioration. Whether or not older patients who already show significant developmental delay should be offered BMT as an option is both an ethical and moral question as well as an empiric one. However, at the present time, it appears that this decision should be based on an expectation that change in developmental function after BMT is similar regardless of pre-BMT levels and that some decline relative to pre-BMT functioning is likely for most patients.
Response
The authors thank Drs Phipps and Mulhern for their thoughtful response to our recent publication on the outcome of unrelated donor (URD) bone marrow transplantation (BMT) in 40 children with Hurler syndrome (MPS I H).1-1 We would like to respond to several important points, including the objection that our conclusion “MPS I H patients with a baseline MDI greater than 70 who are engrafted survivors following URD BMT can achieve a favorable long-term outcome and improved cognitive function” is not warranted. We would like to emphasize that our outcome measure is not MDI but the trajectory of the child's mental development.
To assess the effects of any treatment on outcome in a developing organism one must measure change reliably. Most research studies on change after treatment use difference scores. A major problem in statistical analysis is the unreliability of difference scores. Difference scores require that the measure being used have high reliability and that the interval being measured be similar. As neuropsychologists know, infant tests suffer from decreased reliability due to behavioral and other external factors. Also, in samples such as ours in which the interval between treatment and outcome varies, we need a dependent measure that is sensitive to change within the individual and can be applied when the number of observations varies and some observations may be flawed. We have chosen to study developmental change by examining changes in raw scores or age equivalent scores as a slope rather than as a single end point.1-2 Although our small sample only permits simple conclusions, we can examine the slope of development and not unreliable difference scores. We did not use calculated MDI scores as our outcome measure. Instead, we were determining whether transplanted children continued to develop on trajectories similar to their baselines. Our dependent measure was rate of growth using age equivalent scores, which are similar to raw scores. We wished to compare children's rates of skill acquisition to normative samples. Thus, our conclusion of “improved cognitive function” is a within-child comparison. We were interested in knowing whether they slow, whether they plateau in development, or whether they lose function, as untransplanted MPS I H children inevitably do.
Among the 6 transplanted subjects with MDIs greater than 70, we can be certain that only 2 subjects showed a decline over time in standard scores. A third child who had a pre-BMT MDI score of 114 is likely to have been overrated because of correction for prematurity. No child lost any raw score points and all showed an increase in mental age equivalent scores. Please note that two children whose MDIs were less than 70 actually lost raw score points and that their mental age equivalent declined, while another child showed only a 1-month gain in mental age equivalent over the course of 2 years.
Calculation of “ratio IQs” or developmental quotients presupposes that there is similar variation at each age level. This concept has been abandoned because standard deviations vary at each age level and the range of normal performance also changes. The goal is to move from the concept of developmental quotient or MDI as an outcome variable and to look at the effects of BMT on the development of the central nervous system and cognitive function.
We agree that we have a small sample size that is composed only of those MPS I H children who had URD BMTs. Further results, soon to be reported, in MPS I H children who underwent genotypically identical or phenotypically matched related donor BMT will add considerable neuropsychological outcome data.
Finally, the statement that there is a decline in standard scores over time with transplant is not warranted. A portion of the sample did stay the same or better, and half were in the average range of intelligence at outcome on the Stanford Binet Intelligence Scale or McCarthy Scales of Children's Abilities. We also note that, in the past 2 years, with decreased BMT-related morbidity and mortality and earlier transplant, we are seeing fewer children who are slowing in development after BMT.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal