High-Dose Intravenous ImmuneGlobulin for Stiff-Person Syndrome
All data were analyzed before the study code was broken, accordingto the linear model of Kenward and Jones. We extended thismodel to accommodate eight periods of analysis and obtain unbiasedestimates of any base-line differences in the two groups duringeach treatment phase, the direct effects of treatment on symptomsduring the month after each infusion, the residual effect afterthree monthly infusions (first-order carryover effect),andthe residual effect after the washout period (second-order carry overeffect), with use of least-square techniques. Significance testingwas performed by applying permutation techniques that permutethe data in all possible ways and determine a P value for each.The net differences in the stiffness index and heightened-sensitivityscores from base line to the end of three months of treatmentand the first- or second-order carryover effects were comparedbetween the treatment groups in each period. For the distributionof stiffness, a separate subanalysis of stiffness in the trunk,abdomen, arms, legs, and face was also performed.
We also conducted a sequential analysis of the two carryovereffects(at an level of 0.05). We evaluated data for asecond-ordercarryover effect before evaluating whether there was afirst-ordereffect. Depending on the results of the sequentialtests, themodel could include both carryover effects, only thefirst-ordereffect, or neither carryover effect. For example, if theP valuefor the model that included both carryover effects was notsignificant,the final model would not include any carryovereffects. Thefinal model was ascertained on the basis of thetwo-sided Pvalue for the two carryover effects and was used in theanalysisof the effects of treatment and base-line differencesbetweenthe two groups in each treatment period.
Results
Clinical Observations at Base Line
Of the 16 patients, 8 (4 women and 4 men) were randomly assignedtoreceive placebo for the first three months and 8 (5 womenand 3 men)were assigned to receive intravenous immune globulinfirst. The twogroups were similar with regard to age, durationof disease, age atonset of symptoms, severity of disease, andthe prevalence of other autoimmune diseases . Complete results could not be obtained for two patients andthuswere never entered into the analysis. One man who was assignedtoreceive intravenous immune globulin first had a severe, long-lasting,blisteringrash after each infusion that unmasked blinding.One woman who wasassigned to receive placebo first had an unusualpattern of motorbehavior with remarkable daily fluctuationsthat precluded datacollection.
Efficacy
In the group that received placebo first, the mean distribution-of-stiffnessscores did not change significantly during the three monthsofplacebo administration but decreased significantly duringthe threemonths of immune globulin therapy (P=0.01) . In contrast, the scores in the group assignedtoreceive immune globulin first dropped significantly (P=0.02)duringthe three months of immune globulin therapy, remainedconstantduring the washout period, and then increased duringplaceboadministration but did not return to base-line values.Thedifferences in scores between placebo and immune globulinweresignificant at months 3, 4, 5, 7, and 8 . Whentheoverall changes were compared between the two groups, immuneglobulintherapy was found to have a significant direct treatmenteffect(P=0.01) and first-order carryover effect (P<0.001).Subanalysesshowed that immune globulin therapy significantlyreduced stiffnessof the trunk (P<0.001 for the direct treatmenteffect and P=0.04for the first-order carryover effect), abdomen(P<0.001 for thedirect treatment effect), and face (P<0.001for the first-ordercarryover effect).
The time it took patients to walk 9.1 m decreased significantlyinthe group that received immune globulin first, indicatingreducedstiffness and spasms (P=0.02 for the direct and first-ordereffects;P=0.03 for the second-order carryover effect).
Clinical Observations during the First Three Months of Treatment
Six of seven patients who received immune globulin first wereableto walk more easily or without assistance for the firsttime inmonths or years . The frequency of falls decreased,and their fears about crossing open spaces diminished. Theywereable to appear in public, socialize, cross a street withouthelp,shower without spasms, and assume work-related or householdchores.Their faces became animated. In contrast, no such objectivechangesoccurred during the first three months in the sevenpatients whoreceived placebo first.
Clinical Observations during the Second Three Months of Treatment
The condition of five of seven patients who received placebofirstimproved after they received immune globulin, and thecondition offour of seven patients who received immune globulinfirst worsenedonce they began to receive placebo. The conditionof one patient whoreceived placebo first worsened substantiallyduring the threemonths of placebo treatment and was markedby continuous spasms andstiffness (status spasticus)3 by thetimethe crossover phase began. The life-threatening natureof this stateprompted us to break the code. The patient's conditionimproveddramatically after he crossed over to immune globulin.Overall, 11patients had clinical improvement after receivingimmune globulin,according to observations made before the studycode was broken.
Patients' Own Assessments and Follow-up
Of 14 patients who were contacted after the results were analyzed(thespouses of 2 patients who died were contacted), 12 readilyidentifiedthe treatment phase on the basis of the unequivocalimprovement intheir condition during that time. One patient,who received immuneglobulin first, had a slight improvementthat was maintainedthroughout the study, presumably owing toa sustained carryovereffect, and could not distinguish onephase from the other. Thispatient did not wish to pursue furthertreatment with immuneglobulin. The other patient, who receivedplacebo first, recalledhaving a gradual improvement throughoutthe study.
Two patients died after the study: one from gastrointestinalbleedinga year after the study and the other from cardiac arresttwo yearsafter the study. Both had sought and received immuneglobulintherapy after the study. Of the other patients whosuccessfullypursued continued immune globulin treatment, sevenrequire infusionsevery 5 to 12 weeks and one other requiresthem every 4 months inorder to engage in routine daily activities.Two patients did notneed any additional treatment for up toa year. One patient wasunable to obtain approval for immuneglobulin therapy from theinsurance company.
Anti-GAD65 Antibody Titers
Anti-GAD65 antibody titers were measured in six patients ineachgroup. Anti-GAD65 antibody titers remained stable in patientswho receivedplacebo first, whereas they decreased by 33 percentin patients whoreceived immune globulin first and then reboundedsignificantly (P=0.03) during placebo administration .Althoughafter the first three months, the difference inantibody titers was not significant between the two groups ,thetiters were found to have declined significantly (P=0.05)afterimmune globulin therapy when the data for the second periodwerealso included and compared (data not shown). Weekly determinationsofantibodies in two patients showed that titers began to fallby theseventh day after the infusion and reached a nadir withinthreeweeks (data not shown). The anti-GAD65 antibody titersdid not correlate with the severity of disease, and the reductionin titers was not correlated with the degree of improvementin the patients'clinical condition.