SEARCH

SEARCH BY CITATION

Keywords:

  • diet;
  • fat;
  • lipase inhibitor;
  • obesity;
  • weight loss

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Abstract. Franson K, Rössner S (The Swedish Association for People with Bowel and Stomach Diseases, Stockholm and Huddinge Hospital, Huddinge, Sweden). Fat intake and food choices during weight reduction with diet, behavioural modification and a lipase inhibitor. J Intern Med 2000: 247: 607–614.

Objective. To study the composition of fat intake and fat-rich meals consumed during a trial in which obese subjects were treated with a lipase-inhibitor or placebo, with emphasis on food choices and eating hours.

Design. Patients were instructed to record all food and drink taken for four days prior to each dietician visit. The food diaries from all scheduled 15 treatment visits were analysed for nutritional content and composition and for temporal distribution. All meals containing 25 g of fat were defined as fat-rich.

Subjects. Twenty-eight women and six men, mean age 45.2 ± 10.9 (SD) years with a mean body mass index of 37.3 ± 3.3 (SD) kg m–2 at the beginning of the study.

Results. Fat intake, both as absolute weight and as energy % was generally higher in the placebo group but no significant trend over time could be seen. Fat rich meals were increased by 59% towards the end of the study. Most fat rich meals were eaten at lunch and dinner. Cooking fat, fatty sauces, meat dishes and cheese contributed to the major proportion of fat, both for placebo and drug treated subjects. No major changes were seen in food choice over time.

Conclusion. A lipase inhibitor may affect the amount of fat ingested but does not seem to change major sources of fat. The typical fat-rich meal consumed by these subjects was a meat dish, consumed in the evening.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Inhibitors of pancreatic lipase reduce the intestinal absorption of dietary triglycerides. Portions of ingested fat do not therefore enter the metabolism but are excreted in the faeces. This fat malabsorbtion results in faecal energy losses. The most common adverse events of lipase inhibitor treatment are oily spotting, faecal incontinence and flatus with discharge [1].

In a treatment program where a diet and a lipase inhibitor are combined, patients may lose weight in two different and parallel ways: by adhering to the hypocaloric program and by losing energy as fat in the stools as a pharmacological effect of the drug.

Although early studies failed to find differences in eating behaviour between obese and non-obese subjects with regard to fat or carbohydrate intake [2], positive correlations have been found later between overall fat preference and percentage body fat [3]. Obese individuals selected more servings of food and more foods of high-calorie type in an observation study in a cafeteria [4]; obese subjects derived a greater portion of their energy from fat than did lean subjects [5] and list more fat-rich favourite foods [6].

In a Swedish food survey it was shown that cooking fat, table margarine and fatty sauces were the most common sources of dietary fat in the Swedish population [7].

In subjects with an infrequent meal pattern, excessive weight was more common than amongst adults and teenagers who customarily ate five or more meals per day [8 9]. Bray reported that in a group of massively obese patients, prior to obesity surgery, 25% of the daily energy intake was consumed between 22.00 h and midnight; this was considerably less 6 months after surgery [10]. Recent studies from our group did not reveal any major differences in meal-patterns between obese and normal weight men, even after eliminating obvious underreporters [11 12]. Dreon found no relationship between the number of meals, distribution of calories and percentage body fat in middle-aged, sedentary, obese men, but observed that the percentage body fat correlated positively with total fat intake [13]. Kant did not find any association between body mass index and percentage energy from evening food consumption in women [14].

An ongoing study where fat intake was manipulated both by dietary instruction and by faeces losses, induced by the lipase inhibitor tetrahydrolipstatin (Xenical®, Roche), provided an opportunity to analyse the characteristics of fat intake and the composition of fat-rich meals eaten. Food choices and eating hours were of special interest.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The subjects were all out-patients at the Obesity Unit, Karolinska hospital, who participated in a 2-year clinical multicentre trial with a lipase-inhibitor, Xenical® (Research protocol BM 14119C, Hoffman-La Roche Ltd). Forty-seven patients were randomized and 34 completed this trial; only completers were analysed in this study. Completers were divided into two groups during the first treatment year; active (n = 20) and placebo (n = 14). After re-randomization after 1 year, according to the basic research protocol, the patients were divided into four groups, active–active, active–placebo, placebo–active and placebo–placebo. In the second year 16 patients received active treatment and 18 placebo.

The main study was a multicentre, double-blind, randomized and placebo-controlled trial. All patients were given repeated dietary advice in addition to medication. Only subjects from the Stockholm centre in Sweden were included. Twenty-eight of our subjects were women and six were men, mean age 45.2 ± 10.9 (SD) years and mean body mass index 37.3 ± 3.3 (SD) kg m–2 at the beginning of the study.

Patients came to the Obesity Unit at Karolinska hospital twice a month for the first 3 months and once a month for the following 21 months. The study began with a 1-month single-blind placebo run-in period. Randomization was carried out at treatment week 5 and re-randomization at week 57. All patients were given dietary advice by a dietician at each visit; weight was recorded by the study nurse at each visit. All patients obtained their own personal energy level, designed to produce a weight loss of about 2 kg per month during the first treatment year. The median recommended energy level was 6.7 MJ (1600 kcal). The energy level of each patient was decreased by 720 kJ (300 kcal) after 6 months by means of further dietary advice in order to avoid premature weight stability. During the second year energy levels were readjusted to produce weight stability.

Patients were instructed to record all food and drink taken for four days prior to each dietician visit. The food-diaries from all 15 treatment visits were analysed with the computer software program MATS (version 2.2 e 1992; Rudans Lättdata AB, Västerås, Sweden) which includes the Swedish nutrient database PC-kost [15].

All meals containing 25 g of fat were defined as fat-rich. This limit of 25 g of fat was chosen because preliminary studies with the lipase-inhibitor had shown that adverse events were likely to occur when the diet contained around 40 energy percentage of fat. Since the median energy level in the current patient group was 6.7 MJ (1600 kcal), this implied a level of 25 g of fat or more.

Food items and dishes were grouped according to the Swedish national food tables and the contribution from fat, in percentage, was calculated [15]. The term ‘meal’ is used consequently throughout this paper to mean any food and/or drink taken on the same eating occasion.

Adverse events were not frequently reported to the study nurses. It is possible that patients may have been reluctant to report adverse events to the study nurses, believing that the nurses, who controlled the trial medication supply, might suggest to exclude patients with adverse events. Therefore, a dietician interviewed a subsample of 13 patients about adverse events, and what might have caused them. This interview was performed at the last visit.

Statistical analysis

Results are expressed as mean ± standard deviation. Student’s t-test was applied for paired values and unpaired values (comparisons between groups) and anova was used as appropriate; t-tests were two-sided; a P-value of less than 0.05 was considered statistically significant. The study was approved by the Ethics Committee of the Karolinska Institute and all subjects accepted in writing to participate in the study.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Fat intake

Fat intake in g per day during the first and second treatment year is shown in Fig. 1. Fat intake was higher in the placebo group in eight of the 10 food recalls during year 1 and significantly higher at treatment weeks 7 (P < 0.0005), 17 (P < 0.05), 41 (P < 0.05) and 49 (P < 0.05). During year 2, when the diet had become more liberal towards weight stabilization, the fat intake was higher at all times in the drug treated group, reaching statistical significance at weeks 81 (P < 0.05) and 105 (P < 0.05).

image

Figure 1. Fat intake during 2 years of lipase inhibitor/placebo treatment. Re-randomization after 1 year with a more liberal dietary recommendation, aimed at weight maintenance.

Download figure to PowerPoint

Fat intake for the entire 2-year period, with patients divided into four groups by treatment combination, is shown in Fig. 2. Group 3 (n = 6) contained three men with higher recommended energy level, hence the higher daily fat intake.

image

Figure 2. Daily fat intake in groups with different active drug/placebo.

Download figure to PowerPoint

Fat energy percentage

Fat energy percentage during the first and second treatment year is shown in Fig. 3; it was higher in eight of the 10 food recalls during year 1 and significantly higher in the placebo group at treatment weeks 7 and 41 (P < 0.05). During year 2 the fat energy percentage was higher in placebo patients in four out of seven comparisons, but never reached statistical significance. During the second treatment year when patients were divided into four groups by treatment, the differences in fat energy percentage did not reach statistical significance. Mean energy percentage fat at each treatment visit in the four treatment groups did not change significantly over time.

image

Figure 3. Daily intake of fat during 2 years treatment with lipase inhibitor/placebo treatment.

Download figure to PowerPoint

Energy intake

Energy intake was higher in the placebo group in nine out of 10 food recall comparisons and significantly higher at treatment week 7 and 17 (P < 0.05) during the first year. In the second year the energy intake was higher in the drug treated group on all occasions. During the second treatment year when patients were divided into four groups the differences in energy intake did not reach statistical significance. No significant trend in energy intake over time was seen in any of the groups.

Fat rich meals

The percentage of meals containing 25 g of fat or more in the four treatment groups is shown in Fig. 4. Fat rich meals constituted 9.6% of all meals at the beginning of the study and 15.3% of the meals at the end of the study (P < 0.05). Ten out of 34 patients consumed no fat rich meals at the beginning of the study, but all but three subjects had consumed at least one recorded fat rich meal by the end of the study.

image

Figure 4. Percentage of all daily meals which contained more than 25 g fat in all four treatment groups.

Download figure to PowerPoint

Temporal distribution

The temporal distribution of fat in meals during the day for the whole group is shown in Fig. 5. The majority of the fat was eaten at dinner time, between 16.00 h and 22.0 h and at lunch time, between 10.00 h and 16.00 h; little fat was eaten in the morning or during the night.

image

Figure 5. Temporal distribution of fat rich meals over 24 h over the whole treatment period.

Download figure to PowerPoint

Percentage of fat from different fat sources

The percentage of fat from different sources for the drug treated and placebo groups at the beginning of the study, at re-randomization midway through the study and at the end of the study is shown in Table 1, listed as the 10 most common fat contributors; minor and unsystematic changes seem to occur. Patients on drug treatment consumed fat from similar dietary sources compared to patients on placebo and over time no major shifts were found.

Table 1.  Fat from the 10 major dietary sources in patients treated with a lipase inhibitor before, in the middle, and at the end of the study. Contribution (%) to total fat intake
Week 5Week 57Week 139
 ActivePlaceboActivePlaceboActivePlacebo
Cooking fat, sauces242216191818
Meat and poultry dishes171919211216
Cheese1014891010
Sausage dishes777737
Fish & shellfish dishes656667
Biscuits, cookies, buns533786
Milk, yoghurt334454
Bread554454
Egg dishes333445
Ice cream224142
Cream223363
All other fat sources161522151918
Total100%100%100%100%100%100%

Interviews about adverse events

Thirteen patients were interviewed in detail at the final visit about possible adverse events. Three patients were from group 1, three from group 2, two from group 3 and five from group 4. All but one of the 13 stated that they had experienced an adverse event at some time during the study. The most common adverse experiences were soft stools (seven out of 13), fatty stools/oily spotting (seven out of 13) and flatulence (five out of 13). Of those treated with active drug during the first year three out of six had adverse events often, two answered sometimes and one did not know. Of those treated with active drug during the second year one out of five had adverse events often, two sometimes, two never and one did not know. Of placebo treated patients, two out of seven experienced adverse events sometimes during the first year and one out of seven had adverse experiences sometimes during the second year. The adverse events were mainly abdominal pain, soft stools and flatulence.

Five of 13 patients stated that they had noticed a connection between food items and adverse events. When asked to give examples, butter, cream, fatty pastries, chocolate and salmon were typically mentioned. Most patients (eight out of 13) had not noted any connection between food and adverse events. None of the 13 patients had noted a connection between beverages of any kind and adverse events. One patient had noted adverse events when treated with antibiotics.

Of 13 patients nine guessed correctly what treatment they had had. Amongst those who guessed wrongly it was equally as common to mistake active drug for placebo, as vice versa.

Weight changes

Mean weight loss for the whole patient group after 2 years was 9.0 ± 8.6 (SD) kg, ranging from a weight gain of 4.8 kg to a weight loss of 35.8 kg. Most weight was lost during the first 6 months of the trial, regardless of treatment.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

With the rapidly increasing use of the lipase inhibitor tetrahydrolipstatin (Xenical®) it becomes important to collect information on the practical aspects of the use of this drug.

An obvious drawback with the present study is that the overall multicentre design used food diaries primarily to enhance adherence to the study diet. Thus no pre-trial food data are available. The strength on the other hand is that the dietary evaluation allows a repeated analysis of differences in eating patterns associated with intake of a lipase inhibitor but together with standard dietary advice and a hypocaloric diet, low in fat.

Underestimation

A main problem with this study, as with any trial relying on dietary assessment, is the quality of the food intake data. Since Xenical® is not absorbed there is no way to relate side-effects to blood concentrations. Clearly the behaviour of many patients may have been guided by their perceived opinion about their treatment.

Many dietary assessment methods, including diet records, underestimate habitual energy intake [16]. This is true for normal weight subjects [17] as well as for obese subjects. Lissner calculated the mean daily underestimation to be approximately 2.1 MJ (500 kcal) [18]. Some researchers claim that underestimation is more common amongst the obese [19–21] but this hypothesis has also been challenged [18]. That the problem is a question of underreporting and not undereating has been shown with a doubly labelled water technique by Shoeller et al. [20]. Fricker has demonstrated that small eaters are equally good at, and sometimes even better at, estimating portion sizes and remembering food items than normal or large eaters. Fricker suggests that obese small eaters might be underreporting foods often perceived as unhealthy, such as fats [22]. This may also be true in our study where patients were instructed to cut down on fatty food items specifically. Fat energy percentage is thought to be less sensitive to underestimation bias than total energy intake in obese subjects [23]. Still, in our study, the fact that the mean reported energy intake was 0.7 MJ (165 kcal) lower than recommended at the beginning of the study and 1.7 MJ (395 kcal) higher than recommended by the end is reflected in the weight development of the subjects. This difference is statistically significant in the patient group as a whole, but cannot be reached when patients are divided into four groups. Most weight was lost during the first 6 months of the study in all treatment groups and a certain regain of weight was seen during the second year.

Food diaries may not be representative for any other period of time than the days they comprise. Several patients told us that they had gone through periods of overeating but that the food diaries helped them to straighten up their eating pattern before their next visit to the clinic.

It may be argued that the food diaries mostly pay lip service to the dieticians’ recommendations. However, although dietary advice during treatment year 2, for most patients, was aimed at keeping the energy intake at maintenance level, the food diaries show an increased energy intake.

Dietary advice given to the patients was aimed at keeping the fat energy at 30%. This goal was successfully achieved by the whole patient group on most visits, regardless of treatment group. There may be several explanations for this. One could be the regularity of the study design, monthly visits with a food diary exercise and dietary advice each time. Continuity in the treatment from staff is often claimed by patients to be of importance in weight loss programs; in this study only one dietician and only one study nurse monitored the patients throughout the study. The routine of taking a drug three times a day may serve as a reminder of diet resolutions made and also work as a reminder to eat at least three times a day.

Fat rich meals

The overall impression of the comparisons between drug and placebo groups is that fat intake (g) and fat energy percentage did not systematically differ throughout the study. The tendency towards a more fat rich consumption in the drug treated group in year 2 is difficult to interpret. The shift may reflect a chance finding; alternatively, drug treated subjects may have learnt that they could increase their fat intake without developing adverse gastro-intestinal problems and titrate the amount of fat they could tolerate. The split into the four subgroups ( Fig. 2) does not provide any insight into underlying causes.

Fat rich meals were most common in the afternoon or evening ( Fig. 5). These meals were mostly eaten in the homes of the patients, whereas lunches were more often eaten outside home. To reduce overall fat intake, the patients in this study may benefit from practical lessons in low-fat cooking to practice at home. Figure 5 illustrates that the breakfast meal generally contains less than 10% of the total fat intake. This would suggest that the effect of a lipase inhibitory drug taken with breakfast has minor effects on faecal fat excretion, whereas the main part of the daily fat consumption occurs during lunch hours (around 40%) and at dinner time (around 55%). It is possible that for obese patients, who not uncommonly suffer from a night eating syndrome [24], it might be clinically more meaningful to take the third dose before the recurring night eating episode.

In the beginning of the study the major fat sources were cooking fats, fatty sauces and cooked dishes. As an effect of dietary advice patients seem to have been encouraged to eat more prepared food and less snacks. Some patients were also advised to increase their intake of fat since there was an initial tendency amongst the patients to exclude fat altogether when embarking upon the trial. By the end of the study prepared meals seem to get more competition from cakes, desserts and other sweet tasting items. Whether this is a relapse into a pre-treatment eating pattern is not possible for us to evaluate, since no pre-trial food information is available.

The underrreporting problems in dietary assessment were recently discussed by Tonstad et al.[25] in a study of patients at a lipid clinic. In addition, in this group, where obesity was not the primary target of the dietary intervention, food intake was markedly underrreported and 64% of these patients fell under the suggested Goldberg et al. cut-off value of energy intake over basal metabolic rate [26]. There was also a tendency to report fat intake as was perceived socially desirable, with underrreporting of foods which had been restricted in the diet. Clearly the fat intake reported in our study is lower than in reality. If side-effects were as modest as revealed by the interviews, this would indicate that patients could tolerate quite a high daily intake of fat before clinically significant side-effects appeared, or alternatively that drug compliance in high fat meal situations was reduced.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was supported by the Research Center for Eating Disorders, Stockholm County Council and Hoffman La Roche, Basel.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Received 26 November 1999; revison received 26 October 1999; accepted 1 December 1999.