MedicineLakex

 

Dokterhoe.nl

Acta Haematol 2003;109:163–168 Received: June 28, 2002Accepted after revision: November 21, 2002 DOI: 10.1159/000070964 Monitoring Hyperhydration during
High-Dose Chemotherapy: Body Weight
or Fluid Balance?
A. Manka A. Semin-Goossensb,c J. v.d. Leliea P. Bakkera R. Vosc aDepartment of Oncology/Haematology, bCentre for Clinical Practice Guidelines, and cDepartment ofClinical Epidemiology and Biostatistics, Academic Medical Centre, Amsterdam, The Netherlands Key Words
weight can safely be used as the only parameter for Hyperhydration W Nephrotoxicity W Congestive heart monitoring fluid retention in case of hyperhydration dur- failure W Fluid balance W Fluid overload W Body weight W ing chemotherapy.
Chemotherapy protocols Copyright 2003 S. Karger AG, Basel W Nursing protocols Body weight and fluid input/output are usually moni-tored for checking fluid balance in case of intravenous Registration of both body weight and fluid input/out- hyperhydration during nephrotoxic chemotherapy. The put in order to prevent fluid overload during intravenous reliability of measuring fluid input/output is uncertain.
hyperhydration in the course of high-dose chemotherapy Moreover, this measurement is redundant, complex, la- seems to be a ‘ritual' act. There is no scientific basis for it bour-intensive and represents an occupational hazard and no effectiveness rationale.
for nurses and other health-care workers handling fluids Hyperhydration with large amounts of fluid like saline or body excreta. In a prospective cohort study, we deter- is mainly used in nephrotoxic cytostatic treatments with mined the concordance between body weight and fluid e.g. cisplatin and methotrexate which cause immediate intake/output. We also examined the clinical conse- damage to the proximal and distal tubular cells of the kid- quences with respect to the safety of selecting only body neys [1, 2]. Cyclophosphamide and ifosfamide may cause weight measurement as a parameter for fluid overload.
haemorrhagic cystitis [3]. This nephrotoxicity and blad- A total of 591 combined observations of fluid balances der damage can be prevented by forced diuresis with 4–5 and body weights were collected. We observed a higher litres of saline administered intravenously every 24 h in increase in body weight than in fluid balance. The Pear- order to achieve a minimal diuresis of 100 ml/h [4–7].
son correlation between fluid balance and body weight Even in patients with a normal cardiac and renal func- was relatively low (r = 0.28). With regard to the safety of tion, accumulation of water and salt in the interstitial measuring body weight only, we found 4 cases (0.6%) fluid compartment will occur with hyperhydration. Be- who might not have received furosemide if the fluid cause of the risk of fluid overload and pulmonary oedema, input/output had not been measured, without clinical it is clear that careful monitoring of the fluid balance is consequences, however. After standardization, body 2003 S. Karger AG, Basel Department of Oncology/Haematology F6.155 Fax + 41 61 306 12 34 Academic Medical Centre, PO Box 22700 E-Mail [email protected] Accessible online at: NL–1100 DE Amsterdam (The Netherlands) Tel. +31 20 5666090, Fax +31 20 5669030, E-Mail [email protected] In oncology it is customary to register fluid input/out- between fluid balance and body weight, but failed to find put as well as body weight simultaneously in order to a correlation. Because of a lack of published data, the monitor fluid balance. These controls preferably take Dutch Institute for Healthcare Improvement has based its place several times within each 24-hour period in order to guidelines ‘Sense and Nonsense of the Fluid Balance' on be able to timely observe unwanted changes in fluid bal- consensus and only recommends the use of fluid input/ ance and to be able to intervene, if necessary. Fluid output measurements if it is supported by strong argu- input/output and body weight are registered cumulatively during the entire period of hyperhydration. Above a cer- In light of the uncertainty regarding the policy to be tain cut-off value a diuretic as furosemide is adminis- pursued, we determined the concordance between body weight and fluid balance as parameters of fluid overload There are several objections to this labour-intensive to indicate that body weight and fluid balance are ex- registration. First, it is very likely that these cut-off values changeable. Next, we determined the clinical conse- are based on experience and opinion since there is no evi- quences with respect to the safety of selecting the simplest dence in the literature to support them. Secondly, it is not and most reliable parameter, body weight measurement.
clear how (possibly) divergent fluid balance and bodyweight values should be interpreted. Thirdly, the validityof the measurements is also under discussion: there are Patients and Methods
different views with regard to the registration of fluid Between March and June 2000, all patients treated with cytostat- input/output [8]. It is not clear, for instance, whether and ics and in whom hyperhydration was used were included in a pro- how the intake of soup, fruit, ice cubes, or the occurrence spective cohort study. Patients undergoing high-dose chemotherapy of diarrhoea and vomiting should be registered. Finally, were screened for comorbidity in the out-patient clinic before start- there are doubts about the reliability: measurements of ing this intensive treatment. Patients were recruited at the AcademicMedical Centre (AMC) Amsterdam in the departments of pulmo- fluid input/output are not always performed accurately.
nary disease, gynaecology, and haematology/oncology. Consent from Fluid balance charts are often incomplete and inaccurate the medical ethics committee was not necessary and informed con- [9].Volumes, for instance, frequently need to be estimated sent was not required since no changes in the current policy were and cannot be measured. Since both fluid input/output and body weight are registered cumulatively, the size of Present Situation the error can increase with time.
The AMC employs international and national treatment proto- Another argument for critically looking at fluid output cols. The duration of administration in these protocols varies from 1 is that handling cytotoxic urine of cancer patients is an to 5 days and each treatment course is followed by the next with a occupational hazard for nurses. Studies showed an asso- resting period of at least 1 week. Fluid input/output and body weights ciation between handling cytotoxic drugs and fetal loss are registered during hyperhydration (4 to 5 litres of fluid in each24-hour period) and measured simultaneously 3 times per 24 h. In and/or systemic drug absorption by the health care pro- case of a cumulative fluid balance 1 2 litres and/or a cumulative body vider [10, 11]. Therefore, every possibility to avoid han- weight increase 1 2 kg from the start of treatment 5 mg of furosemide dling of fluids and body excreta is welcome.
is administered.
Body weight measurement also has inherent difficul- ties, but to a lesser degree. The variation in execution, StandardizationIn order for these measurements to be performed as precisely and such as time of measurement, type of scales used, clothing reliably as possible, standardization of ‘body weight measurement' worn by the patient and whether or not the patient has and ‘fluid balance measurement' took place prior to data collection.
urinated prior to the measurement are aspects that need Special attention was paid to standardization of the weight scales to be considered using logistical changes and protocols (type and use) and standardization of the circumstances under which [12]. A quality assurance project analysed the routine the body weight measurements were performed, e.g. time point andfrequency, clothing and shoes worn and prior urination. The results practice of chemotherapy and the role performance of of a recently completed investigation into body weight measurement nurses. One of the conclusions was the need for standard- policy have led to a relatively new standardized protocol [16]. Stan- ization of procedures of measuring body weight [13].
dardization of the fluid balance measurement, e.g. agreement on The sparse literature on this subject does not indicate parameters that should or should not be considered relevant, was whether it is really necessary to register both fluid balance done with the co-operation of dieticians and nutritionists.
parameters, and which parameter would be best in terms Data Collection of measurement error sensitivity and execution simplici- During the study period, all fluid input/output and body weight ty. In 1979, Plaum [14] investigated the concordance measurements registered took place in patients who had been admit- Acta Haematol 2003;109:163–168 ted for a course of treatment with cytostatics involving hyperhydra- Table 1. Basic characteristics
tion. Both medical and nursing patient files were used and data col-lection was performed per patient and per course of treatment.
At the start of each course of treatment, sex, age, diagnosis, comorbidity and data on the treatment (type of cytostatics, treatment duration, etc.) were registered. Every 8 h both body weight and fluid Age, years, mean (range) input/output were registered and the cumulative fluid balance and Fluid balance/weight registrations cumulative increase or decrease in body weight were measured. If Courses/patient (range) necessary, intervening administration of furosemide was also record- Fluid balance/weight registrations/course ed. Possible calculation errors were checked afterwards. Increased of treatment (range) body temperature (1 37.5ºC) or fever (1 38.0ºC), vomiting, and diar-rhoea were registered as well.
The agreement, or concordance, between fluid balance and body Pulmonary oncology weight was determined using the Pearson correlation coefficient for the entire cohort [17]. This designates the magnitude of the relation- ship between these variables. In addition, the Pearson correlation coefficient of the individual first, second, third and fourth fluid bal- Congestive heart failure ance and its corresponding body weight was determined in order to be able to trace specific trends in a possible discordance.
To analyse whether the discordance between fluid balance and body weight increases with the increase in body weight, a Bland- Duration of course of treatment Altman analysis was performed [18]. In this analysis the mean scores 1–2 days, fluid balance ! 7 of difference in body weight minus the mean scores of difference in Middle 3–4 days, fluid balance 7–10 fluid input/output are plotted against the mean scores of difference in 1 4 days, fluid balance 1 11 weight alone. In the Bland-Altman analysis, the difference in bodyweight has been used as a reference value, since this is considered to Type of course of treatment be the most reliable parameter if data are clustered near the zero line, Fluid balance/weight – cisplatin no differences in concordance occur in case of an increase in weight.
Fluid balance/weight – cyclo-/ifosfamide The clinical consequence, in terms of safety, of using only one parameter (body weight) for registration of the fluid balance insteadof both body weight and fluid input/output was analysed in a 2 ! 2table, depicting (dis)agreement between body weight and fluid bal-ance. This way it can be determined how often interventions with with fewer than 7 consecutive observations were per- diuretics had to be applied. In case they had to be applied, whether formed in 24% (143/591). Treatment with cisplatin was this was based on fluid input/output or body weight or both. It gives most frequently administered, namely in 78% (460/591) insight into how many cases with a fluid imbalance one would have potentially missed if only body weight had been registered.
No cases of clinically manifest left- or right-sided con- Finally frequencies of occurrence have been calculated for the fol- lowing factors: vomiting, diarrhoea, fever, calculation errors and per- gestive heart failure were observed. In 1 case, furosemide was administered based on physical findings – the occur- All data were analysed with the statistical package SPSS, version rence of oedematous ankles – but it is unclear whether this incidence actually involved congestive heart failure.
In general, there was a higher increase in body weight than in fluid balance; with a mean difference of 728 mg.
The Pearson correlation between fluid balance and bodyweight of all 591 fluid balances and weight measurements Of 43 patients, 279 person-days were observed. The was r = 0.28. At the start, the Pearson correlation between mean age of these patients (58.1% men) was 45 years all first fluid balances and body weight measurements was (range 18–73). In 91% (39/43) no comorbidity was found.
r = 0.57 (84/591). At the second measurement, r was The patients underwent a total of 84 first and follow-up 0.57(83/591), at the third r was 0.40 (58/591) and at the courses of treatment, in which a total of 591 combined fourth r was 0.46 (42/591).
observations of both fluid balance and body weight (cases) The Bland-Altman plot (fig. 1) shows that the discor- were collected. The number of combined cases with more dance between fluid balance and body weight also in- than 11 consecutive fluid balances (courses 14 days) was creases as the difference in weight measurements in- 70% (416/591). Short courses of treatment (1 or 2 days) creases. This means that if a patient had gained only a Monitoring Hyperhydration during Acta Haematol 2003;109:163–168 High-Dose Chemotherapy Fig. 1. Bland-Altman plot showing that when the discordance between fluid balance and body weight increases, the
difference in weight measurements increases as well.
little weight, his fluid balance was more or less in agree- body weight of one patient appeared to be dramatically ment with his weight, whereas if his body weight had different from the previous and subsequent measurement strongly increased, the discrepancy between fluid balance and must have been a registration error. The other 3 cases and body weight had become much larger.
involved differences between fluid balance and body Next, we investigated the clinical consequence of the weight of 230, 350 and 430 ml/g, in which the fluid bal- concordance between fluid balance and body weight. Of ance remained just 12 litres and body weight barely all included cases, 81% (479/591) showed a balance !2 !2 kg. All 4 cases concerned the first or second fluid bal- litres and !2 kg, which means that no furosemide was ance/body weight registration. The mean age of these necessary. In 1.5% (9/591) both fluid balance and body patients did not differ from the whole group (40 years weight had increased (12 liters and 12 kg, respectively); against 45 years) (table 3).
the administration of furosemide was indicated based on Of the interacting factors, fever, vomiting and calcula- both parameters. In 17% (99/591) the weight increased by tion errors, all occurred relatively infrequently and there- 12 kg, but the fluid balance remained !2 litres, and an fore required no further analysis (table 4).
intervention with furosemide was indicated based onweight increase alone. The percentage of cases with a fluidbalance increase 12 liters and a body weight increase !2 kg was 0.6% (4/591). In these 4 cases, furosemidewould not have been administered if the fluid balance had Considering the fact that there is no gold standard for not been measured (table 2). Upon further analysis of fluid overload, body weight and fluid balance seem to be these 4 cases, involving different patients, the registered logical and practical parameters for monitoring possible Acta Haematol 2003;109:163–168 Table 2. Number of cases above and below
the cut-off level of 2 litres and/or 2 kg ¢Fluid balance 1 2 litres fluid overload in hyperhydration. In this study we investi- Table 3. Four cases: ¢weight ! 2 kg, ¢fluid balance 1 2 litres
gated the concordance between body weight and fluid bal-ance as parameters of (possible) fluid overload in treat- ment courses with cytostatics. We also determined theclinical consequences of only selecting the easiest applica- ble parameter. We found that body weight appears to change more rapidly than fluid balance as a result of fluid administration. The correlation between body weight and fluid balance is rather weak: the maximum correlation is PIN = Patient identification number.
0.57 at the first measurement and decreases to 0.28 whenall measurements are calculated together. The Bland-Alt-man analysis confirms that the concordance decreases asbody weight increases. A possible cause of this discor-dance is the cumulative incidence of error which has been Table 4. Occurrence of possible interacting variables
taken into account in the calculations. Through standard-ization and training, the body weight and fluid measure- ments were assured to be as reliable as possible. It is not expected that more training would have improved theaccuracy of the measurements.
Increased body temperature With regard to the safety of measuring body weight Registrations 1 37.5 ° C
only we found that 4 cases in this study (0.6%) would not Registrations 1 38.0 ° C
have received furosemide if the fluid input/output had Moderate (! 200 cm3) not been registered. Except in 1 case, which was a registra- Severe (1 200 cm3) tion error, the differences between fluid balance and body Calculation errors weight in those 3 cases were so small that they were con-sidered as ‘borderline'. It should be realized that the cut-off points are arbitrary and that if the cut-off value hadbeen slightly increased to 2.5 kg/litre these cases wouldnot have been registered at all. The interesting question is other clinicians would feel uncomfortable if the volume whether the current cut-off value for the intervention, i.e.
status remained unmonitored. Therefore we focused on a administration of furosemide, is too low and needs to be single and effective monitoring parameter.
The sample showed a mix of short and long courses of Patients in our study were relatively young and had lit- treatment, performed in accordance with the current pro- tle comorbidity. So it is not surprising that no case of clin- tocols and with the usual cytostatics, in particular the ically manifest congestive heart failure was observed and nephrotoxic cisplatin. However, the patients were not our means to prevent fluid overload appeared to be ade- selected and therefore can be seen as representative for quate. The question remains whether there is a risk of the oncology patient population in our academic hospital.
right-sided congestive heart failure in noncardiac patients Of course, our results may not be directly extrapolated to treated with hyperhydration. However left-sided conges- other situations involving patients with congestive heart tive heart failure (pulmonary oedema) is a serious compli- failure, for instance in cardiac, nephrologic patients and cation and should be prevented. All in all, we and many especially older patients. But one could also question the Monitoring Hyperhydration during Acta Haematol 2003;109:163–168 High-Dose Chemotherapy effectiveness of using similar parameters to monitor fluid results based on 591 observations a sufficient basis for a overload in those cases. There is one exception to using policy change and the implementation of a new guide- body weight to control fluid overload: this is when pa- tients are bedridden and cannot be weighed. What re-mains very important is the clinical evaluation and theidentification of physical signs of fluid overload by nurses and physicians.
The underlying rationale to opt for body weight only as This study has provided a good argument for only mea- parameter for checking fluid balance is that measuring suring body weight as a parameter for possible flush over- fluid input/output is complex and labour-intensive and it load upon hyperhydration in a course of treatment with is unsure whether it is a reliable measuring instrument.
cytostatics. No longer registering the fluid input/output Inaccurate registration and calculation errors, such as during such treatments hardly has any clinical conse- double notation or omission of fluid input or urine pro- quences and does not affect the patients' safety. Conges- duction, may cause considerable variation in the mea- tive heart failure rarely occurs and clinical parameters surement of fluid balance. Due to the large number of cal- other than body weight, such as oedematous ankles and culations, calculation errors may easily occur. It seems shortness of breath, may also lead to adequate interven- plausible to assume that fewer errors can occur in body tions. The weighing method, with the proper standardiza- weight measurement and that weight is a more reliable tion of procedures, can and should be performed since it indicator to detect potential fluid overload and congestive appears to be reliable, safe, simple and time-saving.
heart failure than fluid balance. With respect to time andcosts no data were found in the literature on the amountof time used for registering and processing fluid input/ output. However, it is clear that omission of fluid balance First we would like to thank the Centre for Clinical Practice registration in chemotherapy protocols will save a lot of Guidelines and the AMC Medical Board for providing financial sup- time. A positive side effect is that the risk of handling port, which has enabled the elaboration of this guideline. We also cytostatic urine incurred by nurses will be much lower. It wish to thank all the nurses and other personnel of the participating has to be said that this is only the case once the weighing wards for collecting data or providing support for this project in procedure has been properly standardized. We found the another way.
1 Safirstein R, Winston J, Goldstein M, Moel D, 7 Jones BR, Bhalla RB, Mladek J, Kaleya RN, 13 Tanghe A, Evers G, Vantongelen K, Paridaens Dikman S, Guttenplan J: Cisplatin nephrotox- Gralla RJ, Alock NW, Schwartz MK, Young R, Van der Schueren E, Aerts R, Lejeune M, icity. Am J Kidney Dis 1986;8:356–367.
CW, Reidenberg MM: Comparison of methods Vermeiren P: Role of nurses in cancer chemo- 2 Schilsky RL: Renal and metabolic toxicities of of evaluating nephrotoxicity of cis-platinum.
therapy administration. Retrospective record cancer chemotherapy. Semin Oncol 1982;9: Clin Pharmacol Ther 1980, 27:557–562.
analysis to improve role performance. Eur J 8 Daffurn K, Hillman K, Bauman A, Lum M, Cancer Care 1994;3:169–174.
3 Goren MP, Wright RK, Pratt CB, Horowitz Crispin C, Ince L: Fluid balance charts: Do 14 Plaum S: Investigation of intake-output as a ME, Dodge RK, Viar MJ, Kovnar EH: Poten- they measure up? Br J Nurs 1994;3:816–820.
means of assessing body fluid balance. Heart tiation of ifosfamide neurotoxicity, hematotox- 9 Chung LH, Chong S, French P: The efficiency icity and tubular nephrotoxicity by prior cis– of fluid balance charting: An evidence-based 15 CBO/VWR: Sense and Nonsense of the Fluid diamminedichloroplatinum(II) therapy. Can- management project. J Nurs Manag 2002;10: Balance. Utrecht, National Organization Insti- cer Res 1987;47:1457–1460.
tute for Healthcare Improvement (in Dutch), 4 Daley-Yates PT, McBrien DC: A study of the 10 Miller SA: Issues in cytotoxic drug handling protective effect of chloride on cisplatin neph- safety. Semin Oncol Nurs. 1987;3:133–141.
16 ter Beek L, Hooijer H: Weegbeleid, ‘Weighing rotoxicity. Biochem Pharmacol 1985;14:38– 11 Baker ES, Connor TH: Monitoring occupation- policy, what's the weight?' (in Dutch). Ex- al exposure to cancer chemotherapy drugs. Am tended essay from the College of Higher Educa- 5 Vogelzang NJ: Nephrotoxicity from chemo- J Health Syst Pharm 1996;15;53:2713–2723.
tion, Training in Nutrition and Dietetics and therapy: Prevention and management. Oncolo- 12 Guenter PA, Moore K, Crosby LO, Buzby GP, the AMC, Amsterdam 2000. gy 1991;5:97–112.
Mullen JL: Body weight measurement of pa- 17 Bland JM, Altman DG: One and two sided 6 Ostrow S, Egorin MJ, Hahn D, Markus S, tients receiving nutritional support. J Parenter- tests of significance. Br Med J 1994;309:248.
Airsner J. Chang P, LRoy A, Bachur NR, Wier- Enter Nutr 1982;6:441–443.
18 Bland JM, Altman DG: Statistical methods for nik PH: High-dose cisplatin therapy using assessing agreement between two methods of mannitol versus furosemide diuresis: Compar- clinical measurement. Lancet 1988;i:307–310.
ative pharmacokinetics and toxicity. CancerTreat Rep 1981;65:73–78.
Acta Haematol 2003;109:163–168

Source: https://www.dokterhoe.nl/uploads/article/0b3d53895bd7e399b1ae4294d3017459cd30606e.pdf