The Use of Biochemical Parameters as Nutritional Screening Tools in Surgical Patients

doi:10.4236/ss.2011.22019

Paper Menu >>

Journal Menu >>

Surgical Science, 2011, 2, 89-94

doi:10.4236/ss.2011.22019 Published Online April 2011 (http://www.SciRP.org/journal/ss)

The Use of Biochemical Parameters as Nutritional

Screening Tools in Surgical Patients

Indraneil Basu, Padmanabhan Subramanian, Matthew Prime, Charlie Jowett, Brian Levack

Queens Hospital, Romford, Essex, RM7 0AG

E-mail: [email protected]

Received January 30, 2011; revised February 27, 2011; accepted Ma rch 20, 2011

Abstract

Nutritional status influences surgical outcome and complication rates. The National Institute of Clinical Ex-

cellence, (NICE) recommends screening patients on admission; yet traditional nutritional screening tools are

underutilised. This retrospective case-control study investigates the association between biochemical factors

and adverse outcomes in orthopaedic patients to ascertain whether they could provide more suitable alterna-

tives to traditional screening tools. 66 patients with fractured neck of femur were investigated. Adverse out-

comes including Length of Stay, (LOS), and deaths were recorded. Total Lymphocyte Counts, (TLC), Serum

Albumin Levels and Haemoglobin levels, were recorded pre-operatively, (pre-op) and post-operatively,

(post-op). Adverse outcomes in those with normal and abnormal biochemical values were compared using

Chi Squared and T Testing. Linear associations were tested for using Pearson rank correlation. Automated

Nutrition Scores Beta, (ANSB) were calculated and their relationship to adverse outcomes investigated. Pro-

tein energy malnutrition was common on admission. However, only 2 patients were nutritionally screened

during admission. Those patients with abnormal pre-op TLC had an increased LOS in hospital. Those with

abnormal albumin and/or TLC had increased mortality rates. Abnormal albumin levels were associated with

a significant 3 fold increase in mortality, (p = 0.009) and post-operative TLC were found to be negatively

correlated with LOS, (r = –0.3, p = 0.038). ANSB were also found to correlate with increased adverse out-

comes although this was not significant. This study demonstrates that nutritional status is poorly assessed on

admission in orthopaedic patients and consequently that provision of nutritional supplements is suboptimal.

This study also demonstrates a highly significant relationship between abnormal albumin and adverse out-

comes and identifies a new correlation between post-operative TLC and LOS. This study confirms that indi-

vidual biochemical parameters and biochemical scores can be used to identify orthopaedic patients at par-

ticular risk of adverse post-op outcomes. These biochemical screening methods may be a more efficient and

reliable way of stratifying malnutrition associated risk on admission.

Keywords: Nutrition, MUST, Mortality, Lymphocyte Count, Albumin, ANS Beta

1. Introduction

Malnutrition in elderly patients has long been recognised

as a predictor of poor clinical outcome. Poor nutritional

status has been identified as a causative factor in sus-

taining fractures [1]. Protein-energy malnutrition has also

been associated with increased post-op complications

including delayed wound healing, infections, decubitus

ulcers and increased mortality [2,3]. Conversely im-

proved nutrition has been shown to protect against geri-

atric trauma [4].

There is a large body of evidence emphasising the

importance of recognising poor nutritional status in hos-

pitalised patients early and addressing it appropriately.

NICE guidelines recommend screening for poor nutri-

tional status on admission facilitating appropriate provi-

sion of nutritional supplementation throughout the ad-

mission. The gold standard of nutritional assessment is

the Subjective global assessment, (SGA) tool which is

both involved and, open to interpretation based inconsis-

tencies. It is infrequently taught during medical educa-

tion and less frequently used on hospital wards. More

simple screening tools are available for use on admission

and are frequently included in admission clerking pro-

90 I. BASU ET AL.

formas. One such tool being the Malnutrition Universal

Screening Tool, (MUST) which highlights those at risk

of malnutrition associated complications. Despite rec-

ommendations and tools being in place, assessments and

screening tools remain notoriously underutilised and

provision of supplementation is generally poor [5]. The

cause for such poor compliance with such tools is un-

clear, but the involved na ture of the validated assessment

tools and the need to frequently involve other healthcare

professionals to screen and assess patients and to provide

supplementation is partly to blame.

Simple biochemical markers including albumin levels

and TLC have long been regarded as markers which can

reflect the risk of malnutrition related complications and

provide simple means of screening for those at risk [2,3,

6-10]. However, use of these markers as indicators of

nutritional status is controversial. There is debate as to

whether abnormal biochemical factors are reflective of

poor nutritional status or whether they simply reflect the

physiological severity of illness. Most acute illnesses

increase the physiological demand on the body and

therefore increase the nutritional requirement of the pa-

tient. Biochemical factors and nutritional status are

therefore intricately related. Complex biochemical as-

sessment tools have been validated and used pre-op to

allocate appropriate nutritional su pport to at risk patients

and have been shown to reduce post-op mortality [11].

Simple biochemical assessment tools have also been

validated against the SGA tool and have been shown to

identify those at risk of malnutrition related complica-

tions [12]. Biochemical factors can therefore reflect

malnutrition associated risk. Whether their role is as a

surrogate or as a direct marker of nutritional status re-

mains unclear. Few validated biochemical systems of

nutritional screening are encouraged in clinical practice

and as a consequence, these important factors are over-

looked in spite of evidence to the contrary. In light of

such poor compliance with traditional screening tools

biochemical markers could provide a simple and effec-

tive alternative which may allow us to identify those at

risk of malnutrition related complications earlier in their

clinical course.

Historically albumin has been the most common bio-

chemical marker used to assess nu tritional status and it is

well known to reflect protein energy malnutrition. Al-

bumin levels have been shown to be predictive of LOS,

in-hospital mortality, and recovery of basic activities of

daily living following hip fracture [4,7]. However albu-

min has a prolonged half-life, (20 days) and serum levels

can be significantly affected by concurrent inflammatory

processes including those caused by surgical interven tion.

This has to some extent precluded serum albumin levels

from being used to monitor in-patient nutritional status.

Pre-admission values however, which have not been dis-

torted by hospitalization and operative intervention retain

their predictive value.

TLCs are also known to reflect protein energy malnu-

trition and surgical outcome. Koval demonstrated that

TLC on admission is predictive for one-year mortality

after hip fracture [4]. Other studies have implicated

Pre-operative lymphopenia as a significant risk factor for

the development of post-operative sepsis and mortality

[6,9]. Therefore TLCs can also act as independent nutri-

tional markers in hospitalized patients.

Both these biochemical factors have been used inde-

pendently to assess risk of adverse outcomes but evi-

dence suggests combining such variables creates in-

creasingly valuable predictors of adverse outcomes [8].

With this in mind several groups have devised composite

scores with demonstrable predictive value. Koval inves-

tigated both serum albumin levels and TLC in surgical

patients and succinctly demonstrated a lower likelihood

of poor outcome in those with normal variables than

those with one abnormal variable. He identified the

highest risk group were those with two abnormal vari-

ables [3]. These results were supported by similar find-

ings by Syeminodis et al. [13]. Brugler went further

combining several biochemical parameters with clinical

parameters to create a six parameter scoring system,

known as the automated nutrition score. The variables

included poor oral intake, occurrence of a wound, mal-

nutrition related admission, serum albumin levels, hae-

moglobin and TLC. Specific cut off values were estab-

lished which were shown to reflect increased risk of ad-

verse outcomes. Brugler confirmed that the screening

tool they created was able to identify those at risk of

malnutrition related complications as previously identi-

fied by the SGA. Brugler also validated an abridged

score using only the biochemical parameters, named the

ANSB score. This system was able to assess patient’s

level of risk based on admission bloods alone. Identify-

ing at risk patients in this manner had the potential to

increase the efficiency of nutritional intervention [12].

Further evidence supporting this method of assessment

was set out by Smith who demonstrated that those scor-

ing 2 and above were at increased risk of developing

significant complications [14].

With a growing body of ev idence to su pport the u se of

nutritional assessments in surgical patients and the ap-

parent ineffectiveness of established assessment and

screening tools it could be beneficial to remind ourselves

of the predictive value of simple biochemical tests. This

study aims to assess the current use of traditional nutri-

tional assessments in orthopaedic practice and to inves-

tigate the relationship between biochemical parameters

and adverse patient outcomes in order to ascertain

I. BASU ET AL.

whether these simple measures could be used as more

effective alternative screening too ls.

2. Methods

One hundred and thirty seven patients, admitted to a

trauma centre over a three month period following frac-

tured neck of femur, were assessed retrospectively at 6

months post-opp. Data was collected from individual

patient records, theatre records, hospital pathology data-

bases and patient admission databases. Biochemical

markers including TLC, Albumin and haemoglobin lev-

els were recorded from the admission blood sample and

the first blood test post procedure.

Those without a fu lly documented length of stay were

excluded. Patients not operated on within 48 hours of

admission were also excluded to prevent in hospital

malnutrition confounding results. Data sets were ex-

cluded if any pre-op parameter investigated was not re-

corded, or recorded from blood samples taken at differ-

ent times. They were also excluded if post-op parameters

were not recorded, or recorded from samples taken at

different times. The normal values for albumin and TLC

were taken from Queens Hospital Biochemical Database

and are were 1.5-4 x109/L for TLC and 35-50 g/L for

Albumin.

After exclusion of data, complete data sets from 66

patients remained. Notes of t he 66 patients w ere analy s ed

for MUST assessments and for provision of nutritional

supplements. Notes were used to establish the adverse

outcomes, LOS and deaths. The measured parameters

and adverse outcomes were processed and analysed us-

ing SPSS 17.

Pearson rank correlation was used to assess linear re-

lationships between age and biochemical parameters and

LOS at the 0.05 significance level. The relationship be-

tween abnormal lymphocyte counts and albumin levels

and adverse outcomes including mortality and LOS were

then assessed using chi-squared and t-testing at the 0.05

significance level. ANSB scores were then separately

calculated using the criteria set out by Brugler et al. [12].

These scores were then compared against adverse out-

comes both independently and grouped within score

categories as set out by Smith et al. [14].

3. Results

Of the 66 patients assessed the average age was 82 years

with 17 males and 49 females, (Table 1). TLC and al-

bumin levels were assessed pre and post operatively and

their relationship with adverse outcomes investigated.

Hemoglobin levels were assessed pre operatively for

later inclusion in ANSB scoring for which pre-op hemo-

globin is a factor along with TLC and albumin. The

mean biochemical values observed and their normal ref-

erence ranges are noted in Table 2. Abnormal pre-op

TLC’s, (Mean: 1.02) indicate that the majority of pa-

tients were nutritionally depleted on admission, (Table

2). Only two patients had been nutritionally assessed

using the MUST and 2 different patients had nutritional

supplements prescribed during their admission.

The adverse outcomes analyzed were mortality and

total length of stay. 20 patients had died at time of follow

up and the average length of stay was 24 days, (Table 1).

Pearson rank correlation showed that age was positively

correlated with length of stay, (Figure 1) and that

post-op TLC’s were negatively correlated with length of

stay, (Figure 2) both findings were significant at the 0.05

level. Those with abnormal pre-op albumin results had a

1 day shorter in hospital stay , (Table 3) but were found

to have a significant 38% increase in mortality , (Table 4).

Patients with abnormal pre-op TLC’s were found to have a

2 day increased LOS, (Table 3) and a 21% increase in

mortality, (Table 4) though these were not found to be

statistically significant. Only one patient was found to

have an ANSB score of 3 and therefore was not thought to

be representative of the score bracket. It can therefore be

discounted from individual analysis. With this result

omitted increasing ANSB scores were associated with a

general increase in mortality and LOS, (Table 5). When

the ANSB scores were grouped, without omissions, into

those scoring between 0 and 1 and those scoring between

2 and 3, the higher scoring group was found to have a

prolonged in hospital LOS being admitted 8 days longer

than the lower scoring group, (Table 6).

Table 1.

Patient Demographics Mean Range

Total Number of Patients 66

Males 17

Females 49

Average Age 82 52-103 , (Years)

Average Length of Stay 23 1-68 , (Days)

Table 2. Mean biochemical values with reference ranges.

Biochemical MeasuresPre-Op Post-Op Normal Ranges

Lymphocytes 1.02* 1.08 1.5-4x109/L

Albumin 38.6 30.4 35-50g/L

Haemoglobin 12.1 12-18g/dL

*Abnormal pre-op lymphocytes suggests protein energy malnutrition on

admission.

92 I. BASU ET AL.

Figure 1. Age against length of stay–demonstrating positive

correlation r = 0.298, p = 0.015.

Figure 2. Post-op lymphocyte counts against length of stay–

demonstrating negative correlation r = –0.256, p = 0.038.

Table 3. Pre-op biochemical markers related to length of stay.

Length of Stay for

Biochemical Measures Normal Values Abnormal Values

Albumin 24.1 +/– 2.44 22.9 +/– 3.72 , (Days)

Lymphocyte Counts 21.9 +/– 5.98 24.1 +/– 2.2 , (Days)

Table 4. Abnormal Biochemical Markers related to Mortal-

ity.

Mortality

Biochemical Measures Alive Dead

Normal 40, (78%) 11, (22%)

Albumin Abnormal 6, (40%) 9*, (60%)

Normal 7, (.88%) 1, (12%)

Lymphocyte Counts Abnormal 39, ( 6 7%) 19, (33%)

*Proportionally higher number of deaths in the abnormal albumin group-

significant at the 0.05% leve l .

Table 5. ANS-B and adverse outcomes.

Mortality

ANS-Beta ScoreAlive Dead+ Length of Stay!

0 70.6% 29.4% 20.7 +/– 3.9 Days

1 68.9% 31.1% 24.4 +/– 2.5 Days

2 66.7% 33.3% 33.7 +/– 13.1 Days

3* 100% 0% 22 Days

*This group had 1 patient within it and can therefore be discounted; +Increase

in mortality associated with increased score; !Increase in Length of Stay

associated with increased score.

Table 6. Grouped ANS-beta scores and length of stay.

ANS-Beta Scores Length of Stay

0-1 23.4 +/– 2.1 , (Da ys)

2-3 30.8 +/– 9.7 , (Da ys)

4. Discussion

Based on pre-operative TLC’s the majority of patients

included in the study were protein-energy malnourished

on admission. In spite of this the majority of patients

were not assessed adequately and they did not receive

adequate nutritional support throughout their inpatient

stay. Although this study is limited to one hospital, cur-

rent literature would suggest that this is not an isolated

finding [5]. Inadequate utilization of standard nutritional

assessment tools is partly responsible for the widespread

failure to recognize those at risk of malnutrition related

complications. Using tools which are more accessible

and less involved could be more appropriate in busy

clinical environments. Alternatively such tests could be

incorporated into pathology databases, creating an auto-

mated assessment tool as originally proposed by Brugler

[12].

Albumin is a well investigated parameter and its role

reflecting malnutrition associated risk is well docu-

mented. This paper demonstrates that abnormal pre-op

albumin can be associated with up to a 3 fold increase in

mortality. This significant finding makes serum albumin

a major prognostic marker at the time of admission iden-

tifying those at significant risk of poor outcome post-opp.

Whether provision of appropriate nutritional supplemen-

tation would reduce this risk is unclear from our data

though other evidence suggests that addressing protein

energy malnutrition with nutritional supplementation can

reduce similar adverse outcomes [15]. This would sug-

gest that pre-op albumin could be used as a surrogate

marker of nutritional status to screen those at risk. Inter-

estingly in this study abnormal pre-op albumin levels

were associated with decreased LOS. This goes against

I. BASU ET AL.

previous evidence but could be attributed to the signifi-

cantly higher mortality observed in the abnormal pre-op

albumin group unique to this study.

This study also highlights a significant relationship

between TLC’s and adverse outcomes. Both LOS and

mortality were found to be higher in the abnormal pre-op

TLC group which supports existing evidence. The linear

relationship between post-op TLC and LOS discovered

in this study however is not well documented in current

research and appears to be a new finding. This study

demonstrates a significant but weak negative correlation

between post-op TLC and LOS. Given the small num-

bers included in this study further investigation with

greater numbers would be required to establish whether a

stronger relationship exists. Post-o p TLC may reflect the

physiological shock sustained during the operation and

may in this way identify those who are less likely to re-

cover quickl y .

ANSB scoring of these patients also correlated with

adverse outcomes, both LOS and mortality. Only one

patient scored 3 using the ANSB system and therefore

figures calculated from this score alone were discounted.

After excluding this patient a clear increase in LOS as

one increase ANSB group becomes evident as does the

increase in mortality. This increase in mortality creates

an overall decrease in survival for those scoring higher

using the ANSB. When the ANSB scores are grouped

into 0-1 and 2-3 as proposed by Smith et al there is a

clear increase in LOS in those from the higher scoring

bracket [14]. Although the numbers used in this study are

insufficient to demonstrate any significant difference

using this scoring system the general trend would sup-

port using the scoring system to stratify risk in orthope-

dic patients. Again this would suggest that such bio-

chemical systems do have prognostic value in assessing

those at risk of adverse outcomes.

In conclusion this study demonstrates that traditional

nutritional screening and provision of nutritional sup-

plementation is inadequate in the majority of patients.

Albumin and TLC however can independently, and as

part of the ANS Beta score, identify those at risk of ad-

verse outcomes specifically mortality rates. The signifi-

cant 3 fold increase in mortality associated with abnor-

mal albumin values and the novel inverse correlation

noted between post-op lymphocyte counts and length of

stay are particularly interesting. Biochemical screening

tools are therefore an effective means of assessing nutri-

tional status. Their ease of calculation also makes them

less time consuming alternatives to clinical assessment

based screening tools. Such biochemical systems could

be integrated into pre-admission assessments or into

biochemical databases to facilitate automated screening

which may improve surgical outcome.

5. References

[1] D. Thiebaud, “Importance of Albumin, 25(OH)-vitamin

D and IGFBP-3 As Risk Factors in Elderly Women and

Men with Hip Fracture,” Osteoporosis International, Vol.

7, No. 5, 1997, pp. 457-462.

[2] B. M. Patterson, “Protein Depletion and Metabolic Stress

in Elderly Patients Who Have A Fracture of the Hipp,”

Journal of Bone and Joint Surgery, Vol. 74, No. 2, 1992,

pp. 251-260.

[3] K. J. Koval, “The Effects of Nutritional Status on Out-

come after Hip Fracture,” Journal of Orthopaedic Trauma,

Vol. 13, No. 3, 1999, pp. 164-169.

doi:10.1097/00005131-199903000-00003

[4] K. J. Koval, “An AOA Critical Issue: Geriatric Trauma:

Young Ideas,” Journal of Bone and Joint Surgery, Vol.

85, No. 7, 2003, pp. 1380-1388.

[5] S. S. Porben, “The State of the Provision of Nutritional

Care to Hospitalized Patients-Results from The Elan-

Cuba Study,” Clinical Nutrition, Vol. 25, No. 6, 2006, pp.

1015-1029. doi:10.1016/j.clnu.2006.04.001

[6] L. J. Grossbard, “Lymphocytopenia in the Surgical Inten-

sive Care Unit Patient,” American Journal of Surgical

Pathology, Vol. 50, No. 4, 1984, pp. 209-212.

[7] M. R. Foster, “A Prospective Assessment of Nutritional

Status and Complications in Patients with Fractures of the

Hipp,” Journal of Orthopaedic Trauma, Vol. 4, No. 1,

1990, pp. 49-57.

doi:10.1097/00005131-199003000-00009

[8] D. T. Dempsey and J. L. Mullen, “Prognostic Value of

Nutritional Indices,” Journal of Parenteral and Enteral

Nutrition, Vol. 11, No. 5, 1987, pp. 109-114.

doi:10.1177/014860718701100517

[9] R. T. Lewis and H. Klein, “Risk Factors in Postoperative

Sepsis: Significance of Preoperative Lymphocytopenia,”

Journal of Surgical Research, Vol. 26, No. 4, 1979, pp.

365-371. doi:10.1016/0022-4804(79)90021-0

[10] M. Lumbers, “Nutritional Status and Clinical Outcome in

Elderly Female Surgical Orthopaedic Patients,” Clinical

Nutrition, Vol. 15, No. 3, 1996, pp. 101-107.

doi:10.1016/S0261-5614(96)80033-1

[11] R. C. Smith and R. Hartemink, “Improvement of Nutri-

tional Measures during Preoperative Parenteral Nutrition

in Patients Selected by the Prognostic Nutritional Index:

A Randomized Controlled Trial,” Journal of Parenteral

and Enteral Nutrition, Vol. 12, No. 6, 1988, pp. 587-591.

doi:10.1177/0148607188012006587

[12] L. Brugler, “A Simplified Nutrition Screen for Hospital-

ized Patients Using Readily Available Laboratory and

Patient Information,” Nutrition, Vol. 21, No. 6, 2005, pp.

650-658. doi:10.1016/j.nut.2004.10.012

[13] P. P. D. Symeonidis and D. Clark, “Assessment of Mal-

nutrition in Hip Fracture Patients: Effects on Surgical

Delay, Hospital Stay and Mortality,” Acta Orthopaedica

Belgica, Vol. 72, No. 4, 2006, pp. 420-427.

[14] R. C. Smith, “An Effective Automated Nutrition Screen

for Hospitalized Patients,” Nutrition, Vol. 25, No. 3, 2009,

I. BASU ET AL.

pp. 309-315. doi:10.1016/j.nut.2008.09.007

[15] P. P. H. Katelaris, G. B. Bennett and R. C. Smith, “Pre-

diction of Postoperative Complications by Clinical and

Nutritional Assessment,” Australian and New Zealand

Journal of Surgery, Vol. 56, No. 10, 1986, pp. 743-747.

doi:10.1111/j.1445-2197.1986.tb02319.x