Importance of dietary advice, nutritional supplements and compliance for maintaining body weight and body fat after hip fracture

Abstract

Objective

Poor nutritional status amongst elderly individuals with hip fractures is well documented. Studies have suggested that 30–50 % of patients admitted to orthopaedic departments suffer from protein-energy malnutrition (PEM).

Design

An 6 month intervention study.

Setting

The study was conducted in Sweden between February 2005 and October 2006.

Participants

Elderly patients with hip fractures (n=32).

Methods

Evaluation of compliance with individual nutritional support and whether body weight and body fat (BF) could be maintained after six months. Evaluation of possible effects of nutritional supplements and dietary advice after hip fracture on BMI, BF, and Mini Nutritional Assessment (MNA).

Results

Overall compliance with supplement intake was 73%. After six months, BMI was unchanged. Women's BF had decreased (P<0.01), although the mean calorie intake with nutritional support was 34 calories per kg body weight/day. Three groups could be identified: one group with increased body weight and BF, one with loss of body weight and BF, and one with increased body weight together with increased TBW and loss of BF. Participants who consumed 0–1 supplements daily lost more weight than those who consumed 2 supplements daily. There was a positive difference (p=<0.001) for women between MNA values at baseline and after six months.

Conclusion

In the present study compliance was satisfactory at the group level, and the energy and protein intake increased significantly. BMI was unchanged during the 6 months period. However, the women lost BF during the study period of with some had increasing total body water (TBW). MNA values for women changed in a positive direction.

Introduction

Illness or handicaps often result in rapid weight reduction, as seen after a fracture, influenza or surgery (1). Poor nutritional status amongst elderly individuals with hip fractures as compared with the general elderly population has been well documented (2). One study showed that 71% of patients with hip fractures were considered to be at nutritional risk due to a BMI below 24 (3). Several studies have suggested that 30-50 % of patients admitted to orthopaedic departments suffer from protein-energy malnutrition (PEM) (4, 5, 6, 7). PEM is also associated with muscular weakness and consequently an increased risk of falling (8). A number of studies have identified various risk factors for hip fractures in the elderly such as low body weight and reduced bone density or osteoporosis (9), low BMI (10), low body fat (11), physical load (10, 12), poor food intake (13, 14) and low intake of vitamin D (14). Weight and BMI alone are not adequate indicators of change in body composition; increasing weight can be due to post-surgery oedema (15) or protein deficiency (16). Development of oedema after surgery is associated with morbidity, and age reduces the ability to excrete administered fluid. Bioimpedance analysis (BIA), which measures the body compartments and not just weight, can potentially identify patients at risk (15). Several studies have shown that dietary supplementation after hip fracture shortens the hospitalisation period (17, 18, 19) and reduces the risk of postoperative complications (18, 19). A meta-analysis performed by the Cochrane Library concluded that there is evidence, albeit weak, in support of the positive effects of oral protein and energy supplements (13). The poor compliance often associated with oral nutritional supplementation was suggested as a reason for the weak evidence (20). Another aspect concerns patients’ perceptions of the quality of the health care they receive. According to one study, factors that directly or indirectly influence patients include empathy, lack of agreement, respect for their pain, and good nutrition (21). The aim of the present study was to investigate compliance with individual nutritional support and to evaluate whether body weight and body fat could be maintained in elderly persons during a six-month period following hip fracture.

Material

The study was conducted in southeastern Sweden, in a city with a population of 134 000 inhabitants, between February 2005 and October 2006. The aim was to include at least 30 patients. The inclusion criteria were that each patient should:

• •have suffered a hip fracture and been treated at the orthopaedic department at the city hospital
• •be living independently
• •be 65 years or older
• •have the cognitive ability to communicate verbally and to read and complete the forms included in the study
• •be able to speak and understand Swedish
• •be capable of standing up on the scales for 30 seconds
• •not have a pacemaker or other internal electrical medical devices implanted, as Bio-impedance Analysis could interfere with the electrical signals

The nurse at the orthopaedic department informed the dietician (Y.W) about patients who were suitable for inclusion according to the criteria. The dietician then met each patient to determine if the patient had the cognitive ability to communicate verbally and to read and complete the forms included in the study.

The dietician asked if the patients were willing to participate and gave them both verbal and written information about the study while they were still in hospital.

Method

Body composition measured by bioimpedance analysis (BIA) and BMI were analysed once each time during the first postoperative week at the orthopaedic department, and then measurements were repeated in the patient’s home after intervals of one, three and six months. Height was measured to the nearest cm with the patient lying on a flat bed.

To determine body composition the BIA leg-to-leg method was used. This is an electrical impedance technique that relies on differences in electrical conductivity between body compartments. With this method a weak electrical signal is sent through the body that different body compartments conduct to varying degrees, depending on their composition (22). The body composition of individual patients was measured with a commercially available BIA “leg-to-leg” instrument (Tanita, TBF- 300, Sunnex Tillquvist AB, Spånga, Sweden). This method determines weight in kg, body fat mass in kg (BF), fat free mass in kg (FFM), body fat in % (BF%), and total body water in kg (TBW). The intra-variation (n=12) on the scale was 0.07% for body weight, 0.39% for BF and < 0.01% for TBW. The scales used in the study were calibrated using a set located at one primary health care centre. BIA validation studies are based on comparisons with gold standard methods for measuring body composition such as water weighing, magnetic resonance and dual-energy X-ray absorptiometry DEXA (23). The Tanita scales have been shown to give good results in elderly persons (24, 25) and in healthy adults (22).

To evaluate the nutritional status of elderly people, the Mini Nutritional Assessment (MNA) was used (26) in accordance with the Swedish MNA User Guide (27), i.e. breakfast, lunch/dinner and dinner/evening meal with each meal containing protein, fat, carbohydrates, minerals, vitamins and trace elements. The MNA values were obtained at baseline in the subject’s home or at the rehabilitation facility immediately after the hospitalisation period, and again after six months. A value below 17 indicates malnourishment and a value between 17 and 23.5 is suggestive of a risk of malnutrition.

During their stay in hospital, each patient received a standard 400 ml of nutritional supplements every day (Novartis, Täby, Sweden). Two different types of energy and protein supplements were used in the study, “Komplett Näring” (Novartis) in three different flavours containing 7.5g of protein and 150 calories/100ml, and “Resource protein” (Novartis) in four different flavours containing 9.4g of protein and 100 calories/100ml.

During their first week and after one month, the dietician visited the patients in their homes, conducted a dietary interview and calculated their oral energy and protein intake. The interviews consisted of a series of questions about appetite and meal patterns, and ended with a 24-hour recall (28). An individual food plan with dietary advice was drawn up for each patient, i.e. the calculated oral and energy intake from the regular diet added to the nutrition supplements needed, for a total of 40 calories per kg/ body weight per day (29, 30). Each patient was given an individualized quantity of nutritional supplements, a total of 200-600 ml each day (1-3 supplements), in their preferred flavours. The amount of protein needed was calculated based on 1.5 g per kg body weight (29, 31, 32). The one-month visit was conducted to check and follow up the calculations at baseline. Possible chronic illnesses of each patient were determined by referring to primary health care medical records.

To enable assessment of compliance with nutritional supplement intake, a list documenting the intake had to be signed each day, either by the patients themselves or by relatives or home help personnel. The dietician collected these lists at each home visit during the whole study period, i.e. after one month, three months, and six months. At each visit the dietician checked the amount of nutritional supplements remaining in the home in order to assess compliance.

Statistics

The statistical software used was SPSS 14.0 for Windows. Changes in groups were analysed with the Wilcoxon signed rank test and Student’s T-test.

Pearson's correlation coefficients (r) were determined in correlation analyses.

Intakes of energy and protein were calculated using the program Dietist for Windows, 1.3 (Kost och näringsdata AB, Bromma, Sweden). P values of <0.05 were considered significant.

Ethics

The Ethics Committee of the Faculty of Health Sciences, Linköping University, Sweden, approved the study (no. M125-04).

Results

Fifty-one persons were asked to participate and 34 of them agreed to be included, of whom 28 were women. Five individuals were excluded by the dietician for lack of cognitive ability to communicate verbally. Common reasons given for refusing to take part in the study were not wanting a stranger to visit them in their homes or a lack of motivation (10 patients), wanting their partner to do the cooking (one patient), and planning to move to another city (one patient). At the time of the study, patients normally spent 10 to 11 days at the orthopaedic department after a hip fracture. Of the 34 patients included in the study, two women died during the intervention. One man (subject no. 32) suffered a stroke two days before the last home visit and could not be weighed on the scales the final time. Twenty-three persons lived alone, of whom 20 were widows/widowers and three were divorced. Nine were married and lived with their partner. Thirteen had home service during the six-month period.

Of the patients with chronic illnesses, 16 had heart disease, four had diabetes and four had been diagnosed with cancer. At the time of fracture, the mean age of the patients was 82 years for women (range 66-95) and 75 years for men (range 66-86). The age range was 66-95 years, and three of the patients were younger than 70 years. The findings for these three younger participants did not have any impact on the results at the group level. None of the patients was operated on again during the study period, although two women had some problems with their implant (No. 4, No. 8). The participants’ cognitive ability to communicate verbally and to read and complete the forms was unchanged during the duration of the study. The mean intervention time after the fracture was 185 days (range 157-209 days).

Dietary advice, nutritional supplements and compliance

An individual food plan, including dietary advice, was designed for each patient. Twenty-two participants received individual dietary advice in the home and 10 at the rehabilitation facility, either individually or in association with the staff at the facility. The mean length of time at the rehabilitation facility was 73 days (range 23-168 days). One patient (no. 6) subsequently moved into a full-care facility.

Based on the dietician’s recommendations the patients who moved into a rehabilitation facility (n=10) or had home service (n=6) received calculated energy and protein-rich food (EP-diet) from the central kitchen, if needed. The most common number of daily nutritional supplements was two (n= 24), but this varied among the participants. Three patients did not receive any nutritional supplements at all. Two of them had a BMI over 25 and an MNA score over 17, and they opted to try to obtain enough energy and protein intake from a regular, unsupplemented diet. The third patient had a low BMI (22.4) but an MNA score over 17, and she also wanted to try to manage her energy and protein intake by herself. One woman (patient no.13) had to discontinue her intake of nutritional supplements due to stomach problems, which she attributed directly to the supplements. She was instead given a supplement-free energy and protein-rich diet for the remaining 42 days of the study period. For the participants who received supplements, intake compliance was 73% for the group as a whole, 75% for women (range 26-100%), and 71% for men (range 66-98%). A difference in weight was found. One group (n=10) had an increase in body weight the first month and had a mean compliance of 83% (range 71-100 %). The other group (n=19) lost weight the first month and had a mean compliance of 67% (range 26-66 %). Both groups had nutritional supplements for a mean of about 3 ½ months, 99 days and 115 days, respectively. There was no difference in compliance between those who received supplements for less than three months and those with supplements for more than three months. Nor were any differences found in the intake of nutritional supplements between those who lived alone and those who were married, or regarding the number of visits from the dietician in addition to the regular visits in the study. If the mean compliance with nutritional supplements of 73% is used as a dividing point between poor and good compliance, 14 participants had poor compliance and 15 participants had better compliance, irrespective of whether they had consumed 1 supplement or 2-3 supplements. At the one-month follow-up after advice from dietician, a mean of 335 calories more per day were consumed at the regular diet compared to the calculated diet at baseline.

Body composition

For the whole group, a significant correlation was found between body weight and BF both at baseline (0.60 (p<0.001)) and at follow-up six months later (0.74 (p<0.001)). Three distinct groups with varying combinations of changes in body weight, BF and TBW could be identified: one group with increased body weight, BF, and BF%, one group with loss of body weight and BF, and one group with increased body weight together with increased TBW and loss of BF (Figure 1). Results for individual body weight in kg, BF, TBW and BMI at baseline, and at one month, three months and after six months are shown in Table 1.

Figure 1.

The difference in body weight, body fat (BF) and total body water (TBW) at baseline and after six months (n=32). The baseline values are illustrated as 0 on the y-axis

Table 1.

An overview of background data together with results for compliance with individual nutritional supplements and body weight in kg, body fat in kg (BF), total body water in kg (TBW) and Body Mass Index (BMI) at baseline, and after 1 month, 3 months and 6 months for the 32 patients with a hip fracture included in the study

Patient no.	Gender Male (M) Female (F)	Year	Weight (kg) baseline (1) 1 month (2) 3 months (3) 6 months (4)				BF (kg) baseline (1) 1 month (2) 3 months (3) 6 months (4)				TBW (kg) baseline (1) 1 month (2) 3 months (3) 6 months (4)				BMI baseline (1) 1 month (2) 3 months (3) 6 months (4)				Nutritional supplement compliance in %
			1	2	3	4	1	2	3	4	1	2	3	4	1	2	3	4
1	F	88	40.1	38.8	37.8	37.5	4.1	1.2	2.2	3.2	26.4	27.5	26.1	25.1	19.1	18.5	18	17.8	84
2	F	87	50.6	51	49.5	49	16.9	10.1	9.9	7.3	24.7	29.9	29	30.5	18.1	18.3	17.7	17.6	36
3	F	85	49.1	48.3	48.5	49.1	13.1	10.1	11.1	10.7	26.4	28	27.4	28.1	22.4	22.1	22.1	22.4	73
4	F	80	41.4	43.7	44.2	45	nd	nd	9.5	9.3	nd	nd	25.9	26.1	17.3	18.2	18.4	18.7	59
5	F	91	60.2	59.9	57.5	52.6	2101	24.5	23.3	17.5	28.6	25.9	25	25.7	24.4	24.3	23.3	21.3	97
6	M	79	66.6	68.4	74.6	72.7	13.1	13.9	16.4	14.5	39.2	39.9	42.6	42.6	23	23.7	25.8	25.2	98
7	F	83	50.4	52.4	54.8	56.1	11.9	12.4	13.4	15	28.2	29.3	30.3	30.1	20.7	21.5	22.5	23.1	91
8	F	85	60.4	58.4	58.1	58.4	17.2	17.6	17.2	18.5	31.6	29.9	29.9	29.2	24.8	24	23.9	24	100
9	F	68	66.5	63.4	65.7	66	23.2	20.8	21	19.6	31.7	31.2	32.7	34	27	25.7	26.7	26.8	100
10	F	95	58.3	56.6	58.8	58.7	nd	nd	20.2	23.1	nd	nd	28.3	26.1	24.9	24.1	25.1	25.1	26
11	F	78	56	53.4	53.9	54.4	16.9	12.7	13	15.4	29.6	29.8	29.9	28.6	22.4	21.4	21.6	21.8	-
12	F	83	76.4	77.6	81.9	83	30.5	24.9	23.9	28.9	33.6	28.6	42.5	39.6	28.1	28.5	30.1	30.5	100
13	F	85	62.3	60	61.5	61.3	21.9	21.2	20.4	21.7	29.6	28.4	30.1	29	25.9	25	25.6	25.5	100
14	M	70	68.8	65.9	67.9	67.6	16.7	14.8	14.9	15.1	38.1	37.4	38.8	38.4	22.7	21.8	22.4	22.3	66
15	F	71	47.7	47.4	49.1	49.9	8.6	3.6	5.5	7.9	28.6	32.1	29.7	30.7	16.9	16.8	17.4	17.7	66
16	F	80	61.7	61	60.6	60.7	19.2	22.1	21.3	21.4	31.1	38.5	28.8	28.8	22.1	21.9	21.7	21.8	95
17	F	83	37.9	34.1	36.5	36.5	5.3	1	3.8	2.3	23.9	34.2	23.9	25	15.8	14.2	15.2	15.2	48
18	F	84	54.9	52.2	52.8	49.1	18.3	15.3	15.3	13.7	26.8	27	27.5	25.9	22	20.9	21.2	19.7	71
19	F	89	53.9	50	49.5	47.6	13.1	14.1	11.5	9.2	29.9	26.3	27.8	28.1	21.1	19.5	19.3	18.6	38
20	M	86	76.8	66.9	69.4	71.5	17.1	15.9	14.7	15	43.7	27.3	40	41.4	25.1	21.8	22.7	23.3	61
21	F	84	61.1	62.7	62.5	61.9	25.6	25.1	25.6	25.4	26	27.5	27	26.7	23	23.6	23.5	23.3	89
22	F	94	50.4	50.2	49.9	50.7	15	7.8	8.6	6.1	25.9	31	30.2	32.7	21.3	21.2	21	24.4	60
23	F	78	46.1	47.3	48.8	47.7	12.2	6	9.2	9.2	24.8	30.2	29	28.2	18.9	19.4	20.7	19.6	57
24	F	83	69.6	70.6	69.6	67	29.8	29.2	29.2	28.2	29.1	30.3	29.6	28.5	26.9	27.2	26.9	25.8	72
25	F	76	81.9	79.6	75	69.7	27.4	21	20.6	17.8	39.9	42.9	29.8	38	27.4	26.6	25.1	23.3	-
26	F	66	43.8	45	48.1	47.5	7.5	9.8	12	12	26.6	25.8	26.4	26	17.8	18.3	19.5	19.3	92
27	F	85	47	52	52	52.3	6.5	10.8	10.8	10.9	29.7	30.2	30.2	30.3	18.4	20.3	20.3	20.4	100
28	M	66	59.1	55.9	63.5	66.4	nd	6.4	9.2	11.2	nd	26.2	29.8	40.4	20.4	19.3	22	23	39
29	F	74	70.3	71	71	69.1	29.6	24.3	25.3	26.3	29.8	34.2	33.5	31.3	27.5	27.7	27.7	27	-
30	F	77	57.5	55.9	55.4	54.9	16.3	14.7	15.5	11.3	30.2	30.2	29.2	31.9	22.5	21.8	21.9	21.4	88
31	M	70	66.8	66	66.3	66.3	12.4	11.2	10.7	11.3	39.8	40.1	40.7	40.3	21.6	21.3	21.4	21.4	95
32	M	82	83.7	81.3	76	nd	13.3	13.8	12.2	nd	51.5	49.4	46.7	nd	25.3	24.5	22.9	nd	66

BF=body fat in kg; TBW=total body water; BMI=Body Mass Index; nd=note done

Values at the group level, for men and for women, for body weight, BMI, BF, FFM and TBW were compared at baseline, and at one, three and after six months (Tables 2a and 2b). After one month there was a significant difference in BF and BF% for women (p<0.01) as compared to baseline (Table 2a), and this finding was the same after both three months and six months. The mean loss of BF for women was 2.2 kg, amounting to almost 4% of total BF. The mean height for the whole group was 161 cm (158 cm for woman and 175 cm for men). There were non-significant differences in weight and BMI.

Table 2a.

Body weight, BMI, BF%, BF, FFM and TBW at baseline, and after 1 month, 3 months and 6 months for women at the group level, mean and range

Variable	Baseline n=26	Range	1 month n=26	Range	3 month n=26	Range	Six months n=26	Range
Weight (kg)	56	37.9-81,9	55	34.1-79.6	56	36.5-81.9	55	36.5-83
BMI	22	15.8-28.1	22	14.2-28.5	22	15.2-30.1	22	15.2-30.5
BF% (%)	29**	10.3-42.8	25**	2.8-41.4	26**	5.7-42	26**	6.3-42
BF (kg)	17**	4.1-30.5	15**	1-29.2	15**	2.2-29.2	15**	2.3-28.9
FFM (kg)	39	32.6-54.5	41	33.1-58.6	40	32.7-58	40	4.4-54.1
TBW (kg)	29	23.9-39.9	30	24.2-42.9	30	23.9-42.5	29	25-39.6
MNA	17***	9-22.5	nd	nd	nd	nd	22.5***	14.5-28.5

BMI=Body Mass Index; BF%= body fat (%); BF=body fat; FFM=fat free mass; TBW=total body water; nd=not done. ** P<0.01; ***P<0.001

Table 2b.

Body weight, BMI, BF%, BF, FFM and TBW at baseline, and after 1 month, 3 months and 6 months for men at the group level, mean and range

Variable	Baseline n=6	Range	1 month n=6	Range	3 month n=6	Range	Six months n=5	Range
Weight (kg)	70	59.1-83.7	67.4	56-81.3	70	63.5-76	69	66.3-727
BMI	23	20.4-25.3	22	19.3-24.5	23	21.4-25.8	23	21.4-25.2
BF% (%)	20	15.9-24.2	19	11.4-23.7	19	14.5-22	19	16.9-22.3
BF (kg)	14	12.4-17.1	13	6.4-15.9	13	9.2-16.4	13	11.2-15.1
FFM (kg)	58	52.1-70.4	55	49.5-67.5	56	53-63.8	55	52.5-58.2
TBW (kg)	42	38.1-51.5	40	36.2-49.4	41	38.8-46.7	41	38.4-42.6
MNA	20.5	16.5-26	nd	nd	nd	nd	24.5	19.5-26.5

BMI=Body Mass Index; BF%= body fat (%); BF=body fat; FFM=fat free mass; TBW=total body water; nd=not done.

Energy and protein

The mean recommendations for calories and protein differed significantly from actual intake at baseline for both women (p<0.001) and men (p<0.03). At the one-month follow-up, the differences between recommendations for calories (p<0.01 women, p<0.116 men) and for protein (p<0.001 women, p<0.03 men) and actual intake were still significant but had decreased (Table 3). The mean energy intake at baseline for women was 24 calories per kg body weight/day and it was 26 calories per kg body weight/day for men. After one month this had increased to 34 calories per kg bodyweight/day for both men and women.

Table 3.

Comparison of mean energy and protein intake with calculated needs for men and women at baseline and after one month

	Women n=26	Range	Men n=6	Range
Energy at baseline (calories)	1334	728-1723	177	1400-2307
Energy needs (calories)	2209	1516-3276	2812	2364-3348
Protein at baseline (g)	49	26-67	68	50-92
Protein needs (g)	84	67-123	105	89-125
Energy after 1 month (calories)	1916***	1196-2482	2372*	1887-2807
Protein after 1month (g)	81***	41-115	103*	84-130

*p< 0.05, *** p<0.001

For the majority of the patients, the actual intakes of protein and energy at baseline were less than the calculated needs. The participants who received 0-1 nutritional supplements a day had a higher baseline calorie intake and a higher body weight compared to those who received 2 or 3 supplements a day (Table 4). After six months the patients in the group with 0-1 supplements a day (n=7) had lost more body weight than the group with 2 or 3 supplements a day (n=25). Both groups had a significantly higher energy and protein intake after one month. The correlation coefficient for both energy and protein intake was 0.91 (p<0.001). The correlation between the difference in calculated protein needs and intake and the difference in baseline and final body weight was 0.54 (p<0.002), and the correlation between the difference in energy needs and intake and the difference in baseline and final body weight was 0.59 (p<0.001). The correlation between the difference in baseline and final body fat weight and the difference in protein needs and intake was 0.44 (p<0.01), and the correlation between the difference in baseline and final BF and the difference in energy needs and intake was 0.46 (p<0.01). At baseline there was a tendency toward a difference (p=0.055) in calorie intake between those who lived alone (n=23, mean 1339 calories) and those who were married (n=9, mean 1618 calories). At the one-month control, however, the calorie intake had increased for both groups, and this difference had decreased, (n=23, mean 1945 calories for those living alone) (n=9, mean 2149 calories for those who were married).

Table 4.

Patients’ body weight, energy and protein intake at baseline and after one month according to the number of nutritional supplements/day

	0-1 nutritional supplements/day n=7 Mean±SD	2-3 nutritional supplements/day n=25 Mean±SD
Energy at baseline (calories)	1503±306	1393±318
Energy after 1 month (calories)	1814±411*	2054±331*
Protein at baseline (g)	51±13	52±14
Protein after 1 month (g)	66±19*	90±16*
Weight at baseline (kg)	67±14	56±10
Weight after 6 months (kg)	62±14	56±10

*p< 0.05

MNA

A difference was found between mean MNA values for women at baseline and after six months, but this was not seen in men (Table 2a, Table 2b). At baseline, 13 women and 1 man were found to be malnourished (MNA<17 points), 13 women and 4 men were in the risk zone for malnutrition (17.5-23.5 points), and one man was well nourished (>24 points). After six months, 3 women were assessed as malnourished and 15 patients (14 women and 1 man) were in the risk zone. In accordance with the Swedish MNA criteria for daily main meals, 7 patients were eating three main meals daily, 15 were eating two main meals daily, and 10 one main meal daily. A significant correlation between BF and MNA was seen both at baseline (0.45, p<0.05) and during the final assessment (0.48, p<0.01).

Ten participants had no contact with a dietician apart from the four visits included in the study, 14 had one or two contacts, and eight had between three and seven external contacts. These contacts were by telephone or in the form of visits to the subject’s home.

Discussion

In the present study compliance was satisfactory at the group level, and the energy and protein intake increased significantly. BMI was unchanged during the intervention, however, there was a significant lost of BF for women. Three distinct groups with varying combinations of changes in body weight, BF and TBW could be identified: one group with increased body weight, BF, and BF%, one group with loss of body weight and BF, and one group with increased body weight together with increased TBW and loss of BF.

This study was undertaken in order to assess the effect on body weight and BF of individually calculated nutritional support after a hip fracture, and to investigate the degree to which it might be possible to achieve compliance with nutritional support. At the time of fracture, the women in the study weighed at least 14 % less than the average Swedish woman and the men weighed at least 6% less than the average Swedish man (33). The study showed that the BF status of elderly patients varied greatly both at baseline and after six months. Most of the participants had decreased BF after six months, although six of them had increased BF. Five of the patients increased in weight; they had lost BF but were found to have increased TBW. This can probably be explained by post-surgical oedema; one patient in the present study suffered from heart failure, which is also associated with increased morbidity (15). Thus it was shown that it is not possible to determine whether a weight increase is due to an increase in BF or in TBW by using only a standard set of scales. This indicates the potential usefulness of Bioimpedance Analysis in nursing homes and hospitals in order to identify patients at risk of either developing or aggravating malnutrition or post-surgical oedema (15). The mean BMI of those investigated was in the lower reference range (BMI<24), indicating that they were already at nutritional risk of fracture at the beginning of the intervention, which is agreement with the findings of others (3). Poor nutritional status in elderly persons with hip fracture is well documented (2), and 30-50% of patients admitted to orthopaedic departments have been found to suffer from protein-energy malnutrition (PEM) (4, 5, 6). Considering the total study group as a whole, as well as the individual patients, no differences in weight or BMI were found between the baseline and the six-month values. Despite persistent efforts by the dietician during the course of the study, it proved impossible to achieve a sufficient energy and protein intake for the participants. At baseline the calorie intake was low, although this had increased after one month to 34 calories per kg body weight/day, which is the minimum necessary to begin to reverse malnourishment (34). This calorie intake seemed to be enough for men, but not for women. Perhaps the higher energy level (40 calorie/kg body weight) used in this study would be helpful for women with hip fractures if they could maintain their BF.

The energy- and protein intake after six months was not documented. However, even when taking into consideration the difference in compliance, the patients who received 2-3 nutritional supplements per day lost less weight during the six months of the study as compared to those who received 0-1 supplement. This indicates that less than two nutritional supplements daily were insufficient to influence weight positively.

The majority of findings in this study regarding nutritional supplements can be attributed to poor compliance (13). Bruce et al reached a similar conclusion and found that only 46% had 100% compliance with all prescribed supplements (20). Our study demonstrated a fairly high compliance rate of 73% for the whole group, although the range was wide. One explanation for the compliance rate found here could be the dietician’s home visits, indicating the importance of being seen regularly and being cared for. Another explanation could be the information given during the visits about the importance of nutritional supplements for healing the fracture. One interesting observation was that patients with increasing body weight also had better compliance. On the other hand there was no difference in compliance between those who lived alone and those who were married, or between those who consumed one supplement and those who consumed two supplements. The study also showed that the most dependent subjects had the lowest compliance. It is reasonable to infer from these findings that the significance of compliance should be kept in mind. Routines should be established and health care personnel should receive training so that dependent patients receive help with nutritional supplements. The aim is to safeguard patients’ safety and quality of life. Thirteen participants had home service during the six-month period, and no information is available about whether the personnel or the patient signed the list for consumption of nutritional supplements. Failure to tailor the individual elderly patient’s meals to his/her needs and wishes was found to be an important issue that needs to be addressed in order to improve compliance (35). Some subjects complained about the large size (200 ml) of the supplements, so one way to increase compliance might be to have smaller sizes. The supplements were protein rich, so more patients received the recommended intake of protein than the recommended intake of energy. This emphasizes the importance of eating three energy-rich main meals, supplemented by two or three light meals or snacks each day. Unfortunately, the participants were generally not able to comply with the Swedish MNA criteria for main meals (27). This illustrates the importance of providing suitable meals.

Conclusion

In the present study, compliance was satisfactory at the group level, but individual compliance varied greatly and was not sufficient in order to reduce the malnutrition of the participants. It was difficult for elderly persons with a hip fracture to consume the recommended intake of energy and protein on a daily basis. BMI was unchanged during the intervention, however, there was a significant difference in BF for women. Three distinct groups with varying combinations of changes in body weight, BF and TBW could be identified: one group with increased body weight, BF, and BF%, one group with loss of body weight and BF, and one group with increased body weight together with increased TBW and loss of BF. Those who best maintained their body mass consumed at least two nutritional supplements a day. There was a positive development in MNA values for women between baseline and the six-month follow-up.

Acknowledgements: This study was supported by the Lions Scientific Foundation of Sweden, and the County Council of Östergötland, and Novartis, for donation of the nutritional supplements. We are grateful to the orthopaedic department for advising us about patients suitable for inclusion in the study, and to the home help personnel and the staff at the rehabilitation facility. Our thanks also go to Rita Ylikivelä for statistical support.