Paradoxical response to sedative/hypnotics in patients with Self-injurious behavior and stereotypy

Paradoxical response to sedative medication has been reported previously among patients with Self-injurious (SIB) and Stereotypic (ST) Behavior (Barron and Sandman, 1983, 1985). The prevalence of this marker was examined in 648 consecutive developmentally delayed patients tested in the EEG laboratory. Several analyses compared response to sedative/hypnotics, assessment of the EEG, and behavioral profiles from the most recent annual comprehensive behavioral/medical evaluation (CDER). The presence of SIB and ST were significantly related to paradoxical response. The EEG was not related to paradoxical response, but had fewer epileptic foci in frequent or severe SIB and/or ST.


INTRODUCTION
Self-injurious behavior (SIB) occurs among 40% of institutionalized psychotic and autistic children (Green, 1967;Shodell and Reiter, 1968) and between 7-19% of mentally retarded individuals (MacKay et aL, 1974;Maisto et al., 1978;Schroeder et al., 1978). Aversive stimulation has been modestly effective in reducing moderate and high-intensity or lifethreatening SIB (Romanczyk and Goren, 1975). However, SIB re-emerges when the patient is withdrawn from these treatment environments (Lovaas and Simmons, 1969;Romanczyk and Goren, 1975  Previous studies suggested that paradoxical responding to sedative/ hypnotic drugs in patients with self-injurious behavior (SIB) and stereotypy (ST) may be a marker of endogenous opiate dysfunction Sandman, 1983, 1985). Patients with SIB and/or ST remained awake and either were resistive, combative, restless, uncooperative, or abusive after administration of sedative/hypnotics. Although the direct mechanism underlying this behavior is not evident from these data, this behavior may be valuable as an index of subgroups of patients with opiate disregulation.
The present study was conducted to extend previous studies of the relationship between SIB, ST, and paradoxical response in a large sample of patients. In addition relationships were included among SIB, other behaviors and neurological status.

Subjects
From a population of 1240 patients, 648 with reliable data for response to sedative/hypnotics and measures of the EEG were included. The general characteristics of the sample are presented in Table I. In this

Procedure
A consecutive series of patients was followed through the EEG laboratory. Patients receiving sedative/hypnotics for performance of the EEG were rated for level of sedation. Ratings were observations made upon arrival at the EEG clinic with the following scale; 1 = light sleep, 2 = wake/sleep, 3 = awake/nervous, 4 = awake/resistive . Dosage levels were not manipulated parametrically but followed the general rules governing clinical practice. Typically, patients were administered the sedative in their residence and were brought to the clinic with postural support. Staff physicians determined probable effective dosage based on weight and history. Although an attempt was made to consider dose/behavior relationships, it was not possible to do so in a reliable manner.
The rating system for the EEG was supervised by a neurologist and is presented in Table II. Blind ratings were made from paper tracings and the degree of abnormality generally reflected the presence of epileptic loci. Other patient behavior was obtained from the most recent Client Development Evaluation Report (CDER) which is completed for each patient annually. Aggression, depression, ST, SIB, hyperactivity, and temper tantrums were rated on a 4-or 5-point scale with the lowest number (1) representing the most severe or frequent rating. Levels of mental retardation, cerebral palsy, hearing and visual impairment were rated with highest numbers representing the most severe or frequent occurrence.

Analysis of Variance
Previous findings were re-examined by comparison of paradoxical response with a 2 (presence of SIB vs. no SIB) • 2 (presence of ST vs. no ST) analysis of variance. A main effect of SIB (F(1,591) = 6.54, p < .01) and ST (F(1,591) = 6.50, p < .01) and their interaction (F(1,591) = 4.64, p < .03) replicated two earlier studies Sandman, 1983, 1985) with much smaller samples. As illustrated in  1, the SIB + ST group had the highest incidence of paradoxical responding. Because age was significantly related to SIB and ST, a subsequent analysis was computed with age as a covariate. The results were virtually identical except that level of significance was slightly lower for the main effects but greater for the interaction (F(1,590) = 5,51, p < .01). A subsequent ANOVA included only patients with high frequency SIB and ST that was judged by clinical staff to be unrelated to environmental influence. Thus, patients were excluded if they exhibited frequent, environmentally-induced (stress) SIB and ST or very infrequent SIB or ST. With age as covariate, the interaction of SIB and ST on paradoxical responding was highly significant (F(1,490) = 8.58, p < .001). As illustrated in Fig. 2, the order of the effects were consistent with the previous analysis.

Correlational and Factor Analysis
Correlations between patient characteristics and SIB severity and frequency are presented in Table III. Severity of SIB was correlated 0.82 (p < .0001) with frequency. Increases in both frequency (r = -0.14, p < .01) and severity (r = -0.17, p < .01) of SIB were significantly correlated with greater incidence of paradoxical responses to sedative/hypnotic medication. Both SIB severity and frequency were related to less severe neurological abnormalities detected in the EEG (r = 0.27 and 0.28, respectively). It is apparent that a large number of variables were statistically (all <0.01) related to both SIB severity and frequency.
Several other findings were of interest. First, because frequency and severity of SIB were highly related, variables that discriminated these behaviors were determined. Depression was significantly (p < .01) related to severity but not frequency of SIB. Conversely, aggression was related to SIB frequency but only marginally to SIB severity. Both severity and frequency of SIB were highly related to frequency of temper tantrums. There was no relationship between sensory loss and ST with hearing loss (r = -.02) or visual loss (r = -.04). The relationship between paradoxical response and EEG was at the chance level (r = -.07).
A Kaiser normalized, factor analysis was computed on a subset of 10 variables of interest with complete data for 648 patients. The variables were age, sex, level MR, number psychotropics, frequency of SIB, severity of SIB, ST, hyperactivity, paradoxical response, and EEG. (The correlational values in this sample remained unchanged). The Principal Components solution was constrained to three factors. The first factor accounted for 24% of the variance and reflected the communalities among severity and frequency of SIB, frequency of ST, the EEG, age, and presence of paradoxical responding (direction of these relations are evident in Table III). The second factor accounted for another 13% of the variance and illustrated the relationship among age, level of MR, and dependence on psychotropics. The third factor explained 11% of the variance and reflected the relationship between hyperactivity and ST.

DISCUSSION
The purpose of this study was to examine the relationship among SIB, ST, and response to sedative/hypnotics. Neurological status of patients with SIB and ST was determined by evaluation of the EEG during sedation. In addition, several demographic variables and behavioral profiles were related to SIB and ST. Reliable relationships were found among SIB, ST and paradoxical responding to sedative/hypnotics replicating earlier findings with smaller samples of patients selected for a higher prevalence of paradoxical excitement Sandman, 1983, 1985). Both frequency and severity of SIB were significantly (p < .01) related to the degree of paradoxical response. These variables, plus ratings of the EEG and age, congregated on the first factor of a Principal Component solution. Factorial modelling, suggested that the presence and rate of SIB and ST, predicted paradoxical excitement. Level of significance was slightly greater when only higher frequency patients were entered into the model.
It remains a curious finding, now repeated four times Sandman, 1983, 1985;Sandman et al., 1990), that paradoxical responding is most prevalent in patients with both SIB and ST. It has been argued that SIB and ST are related, but that SIB is the most maladaptive or regressed expression on a behavioral continuum Cataldo and Harris, 1982). This argument implies that co-occurrence of SIB and ST should be rare, but both often are exhibited by patients. Further, paradoxical excitement is highly prevalent only in patients with both SIB and ST (see Figs. 1 and 2), and patients with both SIB and ST have the highest B-endorphin levels (Sandman et al., 1990). These data suggest that there may be three distinct groups of patients manifesting SIB and/or ST and that the biological mechanisms responsible for the expression of symptoms may be distinct.
The relationship of neurological status to behavior was measured by the level of epileptic abnormality of the EEG. Increasing epileptic abnormality was related to absence of both SIB and ST and there was no relationship between EEG and paradoxical response. It has been suggested that SIB and ST are seizures without recognizable EEG signs (Gedye, 1989). Alternatively, occurrences of epilepsy and SIB or ST may be mutually exclusive. Careful trials with anticonvulsants (see Langee, 1989) in patients with ST and SIB may be warranted even though anecdotal reports indicate they may be ineffective.
Finally, these results can be compared with earlier demographic studies of SIB. The prevalence of daily SIB in institutionalized populations (8.4%) though slightly lower than the Ross (1972) survey of 12%, is within the range of other investigators (Maisto et al., 1978;Schroeder et al., 1974;Smeets, 1971). The estimate of SIB occurring at least three times within a year in 36.5% of our institutionalized population is difficult to compare with previous studies because of differences in the method and criteria of reporting. Sex, level of retardation (though our range was restricted), and length of hospitalization were not related to SIB or ST (see Schroeder et al., 1978). SIB was more frequent in older patients, in patients receiving psychotropic medications, and correlated with aggression, temper tantrums, ST, and hyperactivity. Depression was related to severity but not frequency of SIB (see also Pirodsky et al., 1985). The profile of behaviors was similar for ST, except for the relationship with aggression.
This study confirmed previous relationships among SIB, ST, and paradoxical response to sedative/hypnotics Sandman, 1983, 1985;Sandman et al., 1990). Speculations Sandman, 1983, 1985) connecting this response with the opiate system have received tentative support from direct studies of the opiate system in SIB (Gillberg et al., 1985;Sandman et al., 1990;Wiezman et al., 1984). The ability of opiate blockers to attenuate SIB and perhaps ST (Barrett et al., 1989;Campbell et al., 1988;Herman et al., 1987;Richardson and Zaleski, 1983;Sandyk, 1985;Sandman et al., 1983;Sandman, Barton, and Colman, 1990;Sandman, Barron, Crinella, and Donnelly, 1987) supports the rote of the opiates in this syndrome. Future studies of paradoxical responding will determine whether it is a marker of specific neurochemical and perhaps opiate disregulation in SIB.