Portal de Periódicos da CAPES

Reliability of Height and Weight Estimates in Patients Acutely Admitted to Intensive Care Units

2007; American Association of Critical-Care Nurses; Volume: 27; Issue: 5 Linguagem: Inglês

10.4037/ccn2007.27.5.48

1940-8250

Rogier M. Determann, Esther K. Wolthuis, Peter E. Spronk, Michaël Kuiper, Johanna C. Korevaar, Margreeth B. Vroom, Marcus J. Schultz,

Renal function and acid-base balance

The authors analyze whether estimated height and weight agree with measured height and weight and with values provided by patients or their closest relative.Knowledge of patients' height and weight is essential for daily practice in intensive care units (ICUs). According to current advice,1–3 in patients with acute lung injury or acute respiratory distress syndrome, adjustments in tidal volumes should be based on predicted body weight (which is a function of height). Also, patients' weight is used to adjust the dose of many drugs, and the dosages of infusions of vasoactive drugs are expressed in micrograms per kilogram of body weight per minute so that cardiovascular function can be assessed and comparisons between patients can be made.4 Further, in recent studies,5–11 researchers have investigated obesity as a risk factor for additional morbidity and mortality in critically ill patients.Unfortunately, obtaining patients' height and weight can be difficult. ICU patients are often unable to inform anyone of their actual height and weight. Although patients' height is easily measured by means of a tape measure, in our institutions, height is not measured; rather, it is estimated by the attending nurse. Alternatively, height and weight values can be obtained from patients' relatives, but it is unclear how precise these values are. Furthermore, although weight can be measured by using special beds, the measured weight of acutely admitted patients may not truly reflect the patients' preadmission weight. Indeed, patients in shock may require early infusion of large amounts of fluids, leading to increases in body weight before admission to the ICU.12In this study, we analyzed whether estimated height agrees with measured height and whether nurses' estimates of height and weight agree with height and weight provided by patients and/or patients' relatives. We determined the influence of patients' actual height and body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) on these estimated values. In addition, we investigated whether BMI values calculated on the basis of estimated height and weight values agree with BMI values calculated on the basis of measured height and of weight provided by patients' relatives.Data from consecutive patients acutely admitted to the ICU for whom preadmission height and weight values were unknown to staff were collected prospectively. At admission, a nurse entered each patient's height and weight into the patient data management system (Metavision, iMDsoft, Sassenheim, the Netherlands). For inclusion in our study, the entered values had to be estimates. If a patient's values could be obtained from previous medical charts, the patient was excluded from the analysis. Each patient's height and weight were estimated by eye while the patient was lying horizontally. In addition, the patient's closest relatives (spouses or children) were asked for the patient's preadmission height and weight, and the researcher measured the patient's height with a tape measure. Scales to weigh patients were not routinely available in our institutions. Because acutely admitted patients often require substantial amounts of fluid, we did not weigh our patients. We reasoned that a patient's weight provided by relatives would be a better estimate of the patient's preadmission weight.The protocol was in accordance with the ethical standards of local ethical committees for the protection of human subjects; informed consent was not deemed necessary because of the observational nature of this study and because the study did not require changes in diagnostic or therapeutic strategies.The ICU of the Academic Medical Center (Amsterdam, the Netherlands) is a 28-bed, "closed-format" (ie, care is directed by critical care specialists) mixed department, in which medical and surgical patients (including cardiothoracic and neurosurgical patients) are treated. In the Gelre Hospitals (Lukas, the Netherlands), an affiliated teaching hospital, the ICU is a 10-bed, closed-format mixed medical-surgical department (excluding cardiothoracic surgery and neurosurgery). The Medical Center Leeuwarden, a large general teaching hospital in the north of the Netherlands, has a 17-bed, closed-format mixed ICU (excluding neurosurgery).Mean differences between actual height and provided and estimated heights, between provided weight and estimated weight, and between actual BMI and provided and estimated BMI were calculated. Mean difference was used to assess bias of estimating methods. The 95% limits of agreement were calculated to assess agreement between estimating and asking or measuring methods for individual patients. The 95% limits of agreement are defined as the mean difference plus or minus 2 SDs for normally distributed data and as the limits accounting for 95% of all data for data that are not normally distributed.13 Bland-Altman plots were created to display the individual data on the magnitude of the differences between 2 different definitions of height, weight, or BMI. The middle horizontal line indicates the mean difference. The 2 outer lines indicate the limits of agreement.Finally, to investigate whether estimation was poorer or better in patients with more extreme height or BMI, patients were divided into 4 groups on the basis of actual BMI and 4 groups on the basis of actual height. Differences between these subgroups were evaluated by using the Kruskal-Wallis test. Absolute data are expressed as mean (SD); mean differences are expressed as mean (SE). P values less than .05 were considered significant. All statistical analyses were performed with SPSS version 12.0.2 (SPSS Inc, Chicago, Illinois).In each hospital, 100 patients were included in the study. Height was measured in all 300 patients. Provided height was not available for 6 patients, and provided weight was not available for 7 patients. Estimated height was not available for 3 patients, but an estimate of weight was obtained for all patients. Missing data did not differ significantly among the 3 hospitals.Mean (SD) actual height was 171.4 (10.1) cm (Figure 1). Mean (SD) provided height was 172.0 (10.0) cm. Mean (SD) estimated height was 173.5 (9.9) cm. The mean (SE) difference between provided height and actual height was 0.4 (0.2) cm (P=.006). The 95% limits of agreement were −5.0 to 5.9 cm. The mean (SE) difference between estimated height and actual height was 2.2 (0.4) cm (P<.001). The 95% limits of agreement were −9.9 to 14.3 cm. Most estimated values (96%) were within 10% of the measured values. Moreover, 96% of the estimated values were within 10% of the provided values.The mean (SD) provided weight was 77.7 (15.1) kg (Figure 1). The mean (SD) estimated weight was 78.3 (16.3) kg. The mean (SE) difference between provided weight and estimated weight was 0.9 (0.5) kg (P=.04). The 95% limits of agreement were −14.7 to 16.5 kg. Unlike estimated height values, only 75% of the estimated weight values were within 10% of the provided values.Mean (SD) actual BMI was 26.4 (4.6). Mean (SD) for provided BMI was 26.3 (4.6) and for estimated BMI was 26.0 (4.7) (Figure 2). Mean (SE) difference between actual and provided BMI was −0.1 (0.1) (P=.01). The 95% limits of agreement were −1.8 to 1.6. Mean (SE) difference between actual and estimated BMI was −0.4 (0.2) (P=.02), with 95% limits of agreement of −6.4 to 5.6. When patients were divided into BMI groups, several patients were categorized in a different BMI group when estimated rather than measured or provided values for height and weight were used to calculate patients' BMI (Table 1).Weights of patients with a low BMI tended to be overestimated, and weights of patients with a high BMI tended to be underestimated (Table 2). Moreover, height was increasingly overestimated with increasing BMI. The 95% limits of agreement for estimation of both height and weight were much wider in the highest BMI groups (Table 2). Estimation of height was significantly poorer in the shortest group than in the other groups (P<.001). Height was consistently overestimated in the shortest group compared with the other groups, and the 95% limits of agreement between estimated height and actual height were much wider for this height group (Table 3).We found that estimates of patients' height and weight done by nurses were generally adequate, although misestimations of height and weight were as large as 15 cm and 15 kg, respectively. Height estimates were more accurate than weight estimates. This finding is in line with results in an earlier report14 on estimated height and weight values in critically ill patients. In our study, patients with the lowest BMI were more likely to have an overestimated body weight, and patients with the highest BMI were more likely to have body weight underestimated. Moreover, height was consistently overestimated in the shortest patients. BMI calculated on the basis of estimated height and weight values and BMI calculated on the basis of measured height and provided weight values agreed well. However, several patients fell into the wrong BMI subgroup when estimated rather than measured and/or provided values were used to calculate patients' BMI.Several flaws of the study must be mentioned. First, it is unknown what proportion of the data from the nurses' records was truly estimated. Patients' relatives might have provided height and weight on admission to the ICU. However, we think relatives provided the data in only a few cases, because the patient data management system requires the input of each patient's height and weight directly upon admission, and patients' relatives are usually not present at the bedside at that moment. Also, in the Gelre Hospitals, weight and height had to be entered into a local database, from which dosing tables for vasoactive medications were calculated. In the Medical Center Leeuwarden, height and weight were recorded in a database used for collecting data for the Dutch National Intensive Care Evaluation.15Second, we did not measure patients' weight, and we assumed that weight provided by patients or their relatives would be a better estimate of patients' preadmission weight than measured values would be because acutely admitted patients may have received substantial amounts of fluids before transfer to the ICU. This approach may be an important drawback, because it is unclear how well patients and relatives can recall a patient's weight. Additional studies are needed to address this issue.Third, measuring a supine patient's height with a tape measure may result in different height values than would be obtained by measuring height with the patient standing.16 This bias cannot be avoided in critically ill patients, however.Fourth, increased awareness among our personnel of the importance of the patient's height in applying lung-protective mechanical ventilation may have led to an improvement in nurses' skill in estimating patients' height.17 Therefore, results of our study may not be easily applied to other institutions.The clinical relevance of misestimating height and weight may not be the same for all patients. This situation can be illustrated with some simple examples. A female patient with an actual height of 160 cm has an ideal body weight of 52.4 kg (when the following equation is used: ideal body weight = 45.5 + 0.91[height (in cm) − 152.4]).18 If tidal volumes are adjusted to 6 mL/kg, ideal tidal volumes in this particular patient should be close to 300 mL. Overestimation of height by 14 cm (the higher end of the 95% limits of agreement) results in an increase in her tidal volume of approximately 80 mL, which is roughly 25% of her ideal tidal volume. By contrast, for a female patient with an actual height of 185 cm, the tidal volume adjusted to 6 mL/kg must be close to 450 mL. Overestimation of height by 14 cm also results in an increase in her tidal volume of about 80 mL, which is only 17% of her ideal tidal volume.The same argument applies to overestimation of weight, which may lead to higher doses of medication. We found a systematic overestimation of height, most often in shorter patients. We did not find a systematic overestimation of weight, but the range of the limits of agreement was large. Systematic overestimation in shorter patients has been described before by Coe et al19 but was not found by Leary et al.16 It is unclear whether increases in tidal volume or doses of medication as a consequence of overestimation of height and weight are clinically relevant. For small patients, however, the effect may be important.Not surprisingly, we found a fairly good relationship between actual BMI and provided BMI. For the purpose of this study, actual BMI was calculated on the basis of actual height and provided weight, and provided BMI was calculated on the basis of provided height and weight values. With a good correlation between actual and provided height, these 2 calculated BMI values, by definition, will have a good correlation.For the calculation of estimated BMI, we used estimated height and weight values. In general, estimated BMIs and actual BMIs agreed well. The relevance of misestimations of the BMI by 5 or more depends on what the BMI is used for. If patients are divided into various BMI groups for analysis of outcomes, patients may fall into the wrong BMI group if estimated BMI is used (Table 1). In this case, estimation of BMI may lead to misleading results.We conclude that estimation of height, weight, and BMI by nurses is adequate in general. Height is generally better estimated than is weight. Overestimation of height occurred particularly in the shortest patients, whereas overestimation of weight occurred especially in patients with a low BMI. The consequences of misestimation depend on what height, weight, and BMI are used for. Especially in shorter patients, the consequences of misestimation should be recognized. BMI values differ in important ways, depending on whether they are calculated on the basis of estimated or provided data.Rogier M. Determann and Esther K. Wolthuis contributed equally to this article.