Application Criteria of Enteral Nutrition in Patients With Anorexia Nervosa: Correlation Between Clinical and Psychological Data in a "Lifesaving" Treatment  
From the * Faculty of Psychology, University of
Padua, Italy; and the Correspondence: Agostino Paccagnella, MD, Via Stangade 41, Teviso 31100, Italy. Electronic mail may be sent to apaccagnella{at}ulss.tv.it. Background: Data and research increasingly point to multiple factors in the genesis of eating-behavior disorders, but the lack of a clear etiological definition prevents a unique therapeutic or prognostic approach from being defined. Therapeutic approaches, as well as scientific research, have separately analyzed the psychological aspects and the clinical-nutrition aspects without integrating the variables or correlating clinical and psychological data. This work has several goals because it aims at considering the problem from the 2 different perspectives. Psychological and clinical variables are analyzed both separately and together in order to assess (a) the minimal criteria to define a cure as "lifesaving" and submit a patient to artificial nutrition; (b) the kind of implementation artificial nutrition should follow; (c) which indicators of the efficacy of artificial nutrition must be taken into account; (d) the results in nutrition terms that may be obtained during the follow-up; (e) if artificial nutrition may be used as a therapeutic tool; (f) if there are any psychological effects after artificial nutrition; (g) if there are any effects due to the patients' age; and (h) the correlation between the psychological profile of a patient and the acceptance of the nutrition treatment. Methods: Several psychological and pharmacologic variables, together with clinical and anthropometric data and blood chemical values, were all considered. Conclusions: Besides defining minimal criteria for a "lifesaving" cure and proposing 2 ad hoc scales for the assessment of patients' subjective willingness toward feeding and for the objective measurement of feeding itself, clinical data and correlations with psychological data evidenced the importance of artificial nutrition and specifically of enteral nutrition as a therapeutic tool, allowing us to define the modalities of implementation of enteral nutrition. Results show that, because enteral nutrition did not deteriorate the psychological state of the patients, and was found to be accepted more positively than feeding orally in the most critical initial phase, it should be included in the therapy. Eating disorders (ED) such as anorexia nervosa (NA) and bulimia nervosa (NB) nowadays represent a sociosanitary problem of significant impact in all western countries; NA has an incidence varying between 0.3% and 0.8%,1–4 and is among the psychiatric illnesses with the highest mortality risk (about 10%).5 In order to define a patient as having NA, specific diagnostic criteria both in the clinical and in the psychological field have long been outlined.6,7 NA is considered to have a multiple-factor pathogenesis, in which genetic, psychological, and sociocultural factors all concur, so that in the analysis of the problem and in the planning of the therapeutic approach, not only are predisposing and precipitating factors to be considered but also those factors leading to, maintaining, sustaining a chronic anorectic state.8 and Hence, therapeutic approaches vary in different schools of thought, and several psychonutrition treatments have been implemented in recent decades, some of which tended toward the psychological side, and others toward the nutrition side,9 often failing to combine the variables and correlate psychological and clinical data. The aim of this work is to give a useful contribution in the programming of such treatment by facing the problem of the nutrition approach and discussing it in light of the psychological and physical implications. Among the many aspects we would like to discuss are when is the best moment, in psychophysical terms, to begin artificial nutrition (AN) treatment; which is the best approach to use (enteral or parenteral); and also what behavior is to be taken with patients showing serious periodic physical relapses. Specifically, the questions to answer are (a) what are the minimal criteria to define the treatment as "lifesaving" and to submit a patient to AN; (b) what kind of implementation should AN itself have; (c) what are the indicators of effectiveness of AN which should be considered; (d) which nutrition results may be expected at follow-up; (e) could AN be used as a therapeutic tool; (f) are there psychological effects after AN; (g) are there any effects related to patients' age; and (h) what is the correspondence between the patients' psychological profile and the acceptance of nutrition treatment?
Subjects The study involved 24 female subjects, mean age 18.5 ± 6.18 years, with NA, diagnosed following Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria10; among these, 19 had a diagnosis of restricting-type anorexia, and 5 of binge eating/purging-type anorexia. All the subjects gave an informed consent both to the treatment procedure and to study (which was approved by our Institutional Board for Human Subjects). Because the subjects varied considerably in terms of age, we analyzed such an age difference to assess if it could be an important variable in the evaluation of the subjects, particularly regarding the psychological profiles. Subjects were classified into 3 age categories: 11- to 14-year-old girls (9 subjects), 15- to 20-year-old subjects (8 subjects), and 22- to 32-year-old women (7 subjects). All subjects who took part in the study showed clinical conditions that were considered of "biologic risk" and which justified hospital admission. As clinical literature lacks the criteria to help define such a risk in a clear way,11 the state of "clinical severity" of the subjects and the related lifesaving treatment have been assessed using criteria we set and which are reported in Table I.
At hospital admission (phase I), apart from clinical assessment, routine
anthropometrical
measurements12,13
and the most important blood chemical tests were
performed.14 A
cardiologic examination, an echocardiography, and an ECG were then carried
out. In some cases, specific tests were used to rule out general or
organ-specific pathologies (in 4 cases, an axial cerebral tomography [Siemens,
Emotion 6, Somatom, Erlangen, Germany] scan was performed; in 16 cases,
cerebral magnetic resonance (Siemens, Avanto, Darmstadt, Germany); in all
cases, the tumor markers (ie,
Hospital Treatment In phase II, each patient agreed before hospitalization that treatment would include oral food intake concurrent with enteral nutrition (EN) therapy; willingness to ingest food was subjectively evaluated (Treatment Willingness Scale, TWS; Table II); and actual food intake was objectively estimated. Actual oral intake of food was assessed at the end of each meal through direct observation of food remaining on the plate. To simplify assessment, besides the standard computation of caloric theoretical quota (agreed upon with the patients using the Harris-Benedict equations [HB] computed using the ideal weight = 1336 ± 48 kcal/d19,20), oral intake (evaluated by the WinFood software program, Medimatica, Italy) was classified according to the Objective Caloric Intake Scale (OCIS; Table III), providing a quick comparison with the TWS.
Caloric needs for EN were computed using the HB equation. HB predicted an average need of 1201 ± 41 kcal/d (>30 kcal/kg/d). However, given that in this type of patient the resting metabolic rate (RMR) obtained through indirect calorimetry is typically lower (by about 20%) than the computed value of theoretic metabolism,21 and considering also the clinicometabolic and cardiovascular instability of such patients, it was decided to proceed with an initial caloric infusion of <1000 kcal/d. In phase I, a polymeric formula Nutrison Standard was used (Nutricia: 1 mL = 1.0 kcal; proteins: 19.7%; carbohydrates: 44%; lipids: 36%). In phase II, a concentrated nutrition solution was used (Nutrison Energy, Nutricia: 1 mL = 1.5 kcal). The nutrient was administered by constant infusion for 24 hours. On average, EN lasted 20.7 ± 7.1 days. It was suspended gradually when patients accepted the intake of an oral caloric quota >50% of the theoretical quota before reported.
Psychological Evaluation All 24 subjects taking part in the study, during the 12 weeks preceding hospital admission, underwent a psychometric assessment by means of the self-evaluation questionnaire Self-Report Symptom Inventory–Revised (SCL-90R; scales: Somatization, Obsessive-Compulsive Symptoms, Interpersonal Sensitivity, Depression, Anxiety, Hostility, Phobic Anxiety, Paranoid Ideation, Psychoticism)22; furthermore, 12 of them completed the self-evaluation questionnaire Eating Disorder Inventory-2 (EDI-2; scales: Drive for Thinness, Bulimia, Body Dissatisfaction, Ineffectiveness, Perfectionism, Interpersonal Distrust, Interoceptive Awareness, Maturity Fears, Asceticism, Impulse Regulation, Social Insecurity).23 During hospitalization, a psychiatric and psychological assessment was proposed, together with a psychotherapeutic support, which had the purpose of supporting the nutrition therapy and the hospital stay. At a later point, the treatment became more structured, with a psychoeducational and psychodynamic and family approach. During the hospital stay, the actions of the psychiatrist, the psychologist, and the nutritionist (or the dietitian) were coordinated; treatments did not interfere with each other and could often be carried out in parallel. Family treatment was postponed until the conclusion of the study in order to avoid the addition of confounding factors.
Follow-up Clinical-dietetic control was performed by an expert from our team every 15–20 days in the first 6 months after discharge, and every 30–60 days after the sixth month. Data, which follow, refer to the 12-month follow-up. In 12 patients, it was possible to administer the SCL-90R questionnaire at follow-up.
Statistical Analysis
Clinical and Anthropometric Results Caloric intake (enteral and oral) is reported in Figure 1. At hospital admission, cardiovascular signs were significantly manifest (marked bradycardia: pre = 62.5% vs post = 9.1%, p < .001; arterial hypotension: pre = 66.6% vs post = 27.5%, p < .001; ectopic atrial rhythm: pre = 2.3% vs post = 0%, p < .05; QT prolongation syndrome: pre = 27.4% vs post = 1.2%, p < .001). In the majority, a reduction of the estimated cardiac mass was also present (96%), whereas a slight mitral valve prolapse was recorded in 2.5% of the cases. In Figures 2 and 3 and in Table IV, general symptoms (analyzed only in percentage of the subject population in which symptoms occurred and not by magnitude) and chemical blood parameters are respectively reported. Data show how symptoms were abundantly present at the moment of hospital admission and how they were partly registered even after the nutrition therapy. Some patients who reported persistent constipation during hospital stay were treated with a mix of 3 oils (olive oil = 33.3%; liquid paraffin = 33.3%; almond oil = 33.3%) and blends containing soluble fibers.
Body weight (BW; recorded every 2–3 days and compared with the recommended BMI for NA24) and body mass index (BMI), measured before and after EN treatment and at 12-month follow-up, are shown in Figure 4 (ANOVA: p = n.s.).
The administration of nutrients never caused vomiting, nausea, diarrhea, or a worsened abdominal pain. In a patient (with BMI <11.5 kg/m2) who had a nutrition-metabolic imbalance and showed at admission a bilateral pulmonary effusion, edemas, slight hepatomegaly, or ascites, caloric and volume infusion was operated at an extremely low pace in order not to worsen the general situation. Still, EN reached standard tempo on the 12th day after admission. Two patients (with BMI <11 kg/m2), presenting an edema at the lower limbs at admission, developed in the fourth to sixth day and before standard tempo infusion a picture of minimal refeeding syndrome in spite of the low caloric infusion (respectively: 15.9 and 19.7 kcal/kgactual/d). This was probably the consequence of the high volume of nutrients infused in the course of the 24 hours (respectively 27.4 and 29.2 mL/kg/d).
All the patients at hospital admission reported amenorrhea, were sensitive to cold, and had symptoms of hypoglycemia; by the end of hospitalization, these symptoms disappeared, and only amenorrhea lasted and was recorded in 21 patients (87.5%). In 75% of the patients, the caloric quota ingested corresponded to 62% ± 12% of the theoretic need. A complete normalization of the weight was observed in 6 subjects (25%). As reported by other authors,13 blood chemical parameters remained nearly stable (Table IV), except for a reduction of cholesterolemia (189.75 ± 2.12 mg/dL vs 148.18 ± 7.7 mg/dL; p < .003) and hematocrit (40.00 ± 4.6% vs 36.9 ± 3.6%; p < .002). Six patients who had a cholesterolemia >210 mg/dL at hospital admission were found to have normal parameters at the end of the treatment. Triglycerides were also reduced (p < .001) after EN treatment. Serum electrolytes were slightly altered at hospital admission; in 2 subjects, potassemia was <3 mEq/L. Hemoglobin values at admission were practically within the norm, but they gradually decreased, still remaining within acceptable limits, during hemodilution therapy originating from the introduction of the water and of the nutrient. Values of alanine aminotransferase (ALT) were high in 7 patients and went back to normal values during hospital stay in 5 patients. In the remaining 2, ALT values normalized within 3 weeks after EN was suspended.
Psychometric Results
The EDI-2 questionnaire was administered to 12 patients only. Nine independent sample t-tests were performed between the SCL-90R scores of the 12 patients who answered the EDI-2 and the 12 patients who did not. None of these comparisons yielded significant differences. In a small number of subjects, a 1-way ANOVA was performed on the SCL-90R scale scores across the 3 groups to verify whether there were significant differences due to age. The scores of the 9 SCL-90R scales were the dependent variables, whereas the group was the fixed factor. The main effect of age did not emerge. Similarly, another 1-way ANOVA was performed on the EDI-2 scores across the 3 groups to verify whether there were significant differences due to age. The 11 EDI-2 scale scores were the dependent variables, whereas the group was the fixed factor. Again, there was no main effect of age. No differences emerged in these inventories despite the heterogeneity of subjects' ages (15–32 years). The mean score of the TWS and of the OCIS was calculated for the 3 groups together in the 3 treatment phases. A repeated ANOVA measurement, with age and EDI-2 as fixed factors, was performed on the TWS mean scores (ie, the summary score of the 3 groups in the 3 treatment phases). The TWS scores were significantly different between subjects, [F(2,36) = 3.36, p = .04, p < .05]. A particularly significant difference between phase II and phase III emerged from simple contrasts [p = 0, p < .05] (Figure 7). A repeated ANOVA measurement, with age and EDI-2 as fixed factors, was run on the TWS mean scores of each group separately in the 3 treatment phases. A significant difference emerged between subjects of group 2 (15–20 years) [F(2,14) = 4.93, p = .02, p < .05]. From simple contrasts, it emerged that phase II and phase III differed significantly [p = .02, p < .05]. Only the 15- to 20-year-old group showed significant changes in the TWS (ie, willingness toward feeding increased from phase II to phase III). Multivariate linear regression analyses were performed to verify the effect of the symptoms measured by SCL-90R and of the EDI-2 on the TWS scores. The first linear regression analysis tested the effect of symptoms like Somatization, Obsessive-Compulsive Symptoms, Interpersonal Sensitivity, Depression, Anxiety, Hostility, Phobic Anxiety, Paranoid Ideation, and Psychoticism on willingness toward feeding. The 9 SCL-90R scales were the independent variables; the TWS score in the 3 treatment phases was the dependent variable. Anxiety (ie, symptoms like uneasiness, nervousness, tension, etc) turned out to affect significantly the TWS score of phase I [F(1,23) = 27.59, p = .00, p < .05; R2 = 0.55]; Obsessive-Compulsive symptoms (ie, thoughts, impulses, and actions experienced as overwhelming, etc) resulted to affect significantly the TWS score of phase II [F(1,23) = 34.21, p = .00, p < .05; R2 = 0.60]. Anxiety significantly affected also the TWS score of phase III [F(1,23) = 8.39, p = .00, p < .05; R2 = 0.27]. Another linear regression analysis was performed to test the effect of symptoms like Drive for Thinness, Bulimia, Body Dissatisfaction, Ineffectiveness, Perfectionism, Interpersonal Distrust, Interoceptive Awareness, Maturity Fears, Asceticism, Impulse Regulation, and Social Insecurity on willingness toward feeding. The 11 EDI-2 scales were the independent variables; the TWS score in the 3 treatment phases was the dependent variable. Interoceptive Awareness symptoms (ie, confusion and uncertainty in recognizing and answering in a precise way to emotional states and uncertainty in identifying visceral sensations tied to hunger and satiety) resulted in affecting significantly the TWS score of phase I [F(1,11) = 11.24, p = .00, p < .05; R2 = 0.72]. No significant effect concerning phase II scores. Interpersonal Distrust symptoms (ie, sense of alienation and reluctance in forming close relationships, reluctance in expressing thoughts and feelings to others) turned out to significantly affect the TWS score of phase III [F(1,11) = 9.00, p = .01, p<.05; R2 = 0.47]. Eleven patients agreed to complete the SCL-90R questionnaire even after the treatment (follow-up). For these subjects, 9 paired sample t-tests were calculated between the mean scores of the 9 SCL-90R variables before and after the treatment. No significant differences emerged.
Despite reports that the use of drugs in the treatment of NA may give poor results,25,26 15 patients underwent pharmacologic therapy. Among these, 8 were treated with serotoninergic antidepressants, 3 were treated with dopaminergic antidepressants at low dosage, and 4 with benzodiazepine for the presence of associated hyperactivity and anxiety.
EN and Psychological Aspects Whenever possible, EN should be considered as a part of a therapeutic course focused on a more general change of the person undergoing treatment and certainly not aimed at changes in weight only. In other words, it cannot stand as an occasional or empirical treatment without any logical connection to other treatments. Hospitalization is advised and hence proposed to patients with particularly serious acute medical events.27 As can be seen in the Results, although the subjects in our study were not numerous and were heterogeneous with regard to age and medical profile, they did not differ significantly and, considering the description of the psychological aspects, can be studied as a homogeneous group. In our research, when analyzing willingness toward feeding in the various phases of the treatment, we see how a few days after beginning EN therapy, all the patients continued to be anxious at the thought of gaining weight, although they accept this event in a passive manner, emphasizing the concepts of "clinical necessity" and "lifesaving treatment," and relating them to their hospital stay. Evidence of this is that the highest degree of collaboration (ie, "willingness" toward oral food intake), particularly in the youngest patients, occurs in the late part of hospitalization (10th–15th day of EN, phase II), that is, in a period when the life risk from a clinical point of view is minimal, but while psychologically the risk is high. Anxiety influences willingness toward feeding both in the initial phase of treatment, when the lifesaving procedure, as well as hospital admission and control of the nutrition behavior, are proposed, and in the final phase when the subject has to go back to her daily habits. Other psychological variables influence patients' motivation and are related, in the initial phase of treatment (phase I), to their difficulty in recognizing and responding accurately to emotional states and in identifying visceral sensations (interoceptive awareness). This could explain why these patients have great difficulty in accepting nutrition therapy orally and show a greater readiness toward hospital treatment with EN. In phase II, the variables related to obsessive thoughts and impulses (obsessive-compulsive scale) support other research where the presence of obsessive-compulsive traits related to food intake was evidenced,28 and in phase III where uncertainties and fears regarding relationships (interpersonal distrust) was evidenced. These results stress the importance of agreeing upon a "therapeutic contract," encompassing a slow personalized evolution of the nutrition treatment.29 In particular, our results seem to confirm how high-risk patients (phase I) accept EN better than feeding orally, so that it often becomes an essential therapeutic step for NA treatment.30 In phases II and III, food intake increases in spite of a relatively modest willingness, and data seem to mirror some hypotheses of those who studied patients undergoing enforced nutrition and described phenomena of a disposition toward the food, attributing it to emotional factors.31 On the other hand, there seem to be strong analogies with healthy subjects submitted to semistarvation for several weeks who develop uncontrollable binge eating in the refeeding phase32 or with NA patients who overeat during and after weight gaining recovery.33,34 Our data also confirm the results of Bufano and colleagues,35 who noted in their series of cases a positive effect of EN on the ability of food intake orally. As for the possibility that EN could negatively influence the psychological state of the patients,36 no simple or sure methods have been described in literature to date, meaning there is no definitive conclusion to this problem. Data at follow-up of the SCL-90R inventory show that there are no significant differences in the pre- and post-treatment descriptive psychological profiles of the patients. This shows that the nutrition treatment alone does not significantly improve the psychological state of the patients, but it does not deteriorate it either. EN remains only a therapeutic tool or a means to lifesaving, to activate and discontinue in concomitance of serious clinical or psychological changes. Yet, because temporal observation was limited to a 1-year follow-up, such a result cannot have any prognostic value, given that substantial improvements from the symptoms or full recovery have described at follow-ups been of 5–20 years.37–40
EN and Effects at the Nutrition Level As for the modalities of EN administration, our study confirms other data already available in the literature23: EN can be considered as an important therapeutic tool because it allows increase in the weight of an anorectic patient without the risks usually met with parenteral nutrition (PN; ie, hypophosphatemia, increase in serum transaminase, venous catheters sepsis, etc). An interesting observation regards how a weight increase (and subsequent rise in BMI) does not seem to relate to the weight evaluated at the beginning of the EN nor to the amount of weight loss before admission. Such an increase in weight, which some authors think is the result of a lower energetic expense due to the reduction of agitation with the purging methods, in our opinion comes mainly from liquid expansion, typical of artificial enteral or PN, especially in the initial phase. Indeed, as already described in anorectic patients undergoing a PN,47 this fact appears strongly related, even in our cases, to the reduction in value levels of hemoglobin and hematocrit. For similar reasons, and for the effects derived from a boost in protein synthesis, total protein and albumin level resulted in the norm at the beginning and at the end of EN. Starting from the 10th day of treatment in particular, weight increase might have been mainly due to the increase of food intake (EN + meals orally). This is demonstrated by weight stability maintenance both in the phase of nutrient concentration, and after the end of EN, until hospital discharge and later. It should be noted how weight increase through EN, which must be treated as a basic goal on a global therapeutic plan, has to be included in the therapeutic contract defined before admission. However, the targets should not be to ambitious (eg, 1 kg a week), risking physiologic imbalances, which only demonstrate an inadequate nutrition culture and which are considered dysfunctional, as reported by Gentile.48 In practice, admission for EN and related goals should not create the already-seen gap49 between the physicalorganic and the psychological recovery. On this postulate, initial caloric infusion was programmed at <20 kcal/kgactual/d, gradually progressing toward 30 kcal/kgactual/d, when the initial phase of reduced energetic need is substituted by the refeeding hypermetabolic phase.50,51 As for the 3 cases of refeeding syndrome, where symptomatology appeared although caloric induction was beneath the standards proposed by the majority of the experts, they are just proof that EN must be applied only to selected patients and dealt with by specialized personnel.
Data presented and analyzed in this abstract, although needing confirmation from other thorough clinical-psychological studies, lead us to state that EN should be taken into account even before a case reaches the cited criteria for a lifesaving admission. Data also demonstrate how EN, without deteriorating the psychological state and being better accepted than the food intake orally in the most crucial initial phase, could be inserted within a therapeutic practice. However, it should be adapted and personalized, considering the needs of each single patient, and adequately coupled with the most appropriate psychological or pharmacologic therapy. By considering clinical data and their correlation with psychological data, we could evidence the importance of EN as a therapeutic tool, allowing us to define the minimal criteria of intervention and the modalities of EN implementation. Lack of age-related effects means that the results of this research can be applied to various typologies of subjects, heterogeneous as for the characteristics of their medical problem. Data coming from this research also indicate the need for more studies on a wider range of samples and a deeper understanding of the relation between psychological and clinical variables. Received for publication December 14, 2004. Accepted for publication January 4, 2006.
Journal of Parenteral and Enteral Nutrition, Vol. 30, No. 3,
231-239 (2006)
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
METHODS
RESULTS
-fetoprotein, carcinoembryonic antigen)
were dosed to rule out neoplastic pathologies; in 12 cases, a thorax x-ray
scan was also executed). In all these patients, such tests were repeated at
discharge and at follow-up. 
