Methicillin-Resistant Staphylococcus Aureus ( MRSA ) Isolation from Pharyngeal Swab Cultures of Japanese Elderly at Admission to a Geriatric Hospital

The isolation rate of the methicillin-resistant Staphylococcus aureus(MRSA) from pharyngeal swab cultures in Japanese elderly was studied at admission to a geriatric hospital. The subjects were 233 consecutive patients admitted to Kitakyushu Tsuyazaki Hospital from April 1994 to March 1996. The isolation rate of MRSA was 3.0% in the patients admitted from their own homes , 9.7% in those transferred from nursing homes and 14.0% in those transferred from other hospitals. The patients from their own homes were younger than those from nursing homes, the latter being older than those transferring from other hospitals. The patients from their own homes had better activities of daily living(ADL), higher levels of hemoglobin and serum albumin than those from nursing homes or other hospitals. The white blood cell counts, and the proportion of patients with positive c-reactive protein or with fever did not differ among the three groups. Multiple logistic regression analysis revealed that fever and ADL disability were independent risk factors for the isolation of MRSA, and hypoalbuminemia was a risk factor for MRSA isolation in the model using serum albumin instead of ADL score. These results suggest that the lower isolation rate of MRSA among patients from their own homes may be partly due to better ADL and nutritional status compared with those from nursing homes or other hospitals. J Epidemiol, 1997 ; 7 : 167-172.

then, similar strains have been isolated in other parts of the world Z 3), including Japan 4).Since MRSA represents a high level of resistance for all antibiotics except for a few such as vancomycin and arbekacin in Japan 5), the development of MRSA strains has become a serious clinical and social problem as a causative pathogen of nosocomial infections.Elderly patients have been reported to be one of the high risk groups for MRSA infection 6).It is, therefore, an urgent problem to reduce the incidence of MRSA infection in geriatric hospitals.
Our previous studies showed that disability in the performance of activities of daily living (ADL) 7, 8, 9.11) , poor nutri-tional status such as hypoalbuminemia 8 ,9,10,11) and the use of antibiotics 7.8-9-10,11) are the important risk factors for MRSA infection in the elderly, and that inpatients of a geriatric hospital had a greater positive rate of MRSA than outpatients or the residents of nursing home 7) .Thus far, we have evaluated the risk of MRSA infection in the elderly but not the risk of MRSA isolation 7.8.9.10,11) .Recently, we have studied the prevalence of MRSA isolation at admission in the elderly12).Because of the small number of subjects, however, we could not evaluate the risk factors of the MRSA isolation.The present study thus aimed at evaluating the risk of MRSA isolation among patients at admission to a geriatric hospital.

SUBJECTS AND METHODS
During the two fiscal years from April 1994 to March 1996, 233 consecutive elderly patients of a geriatric hospital routinely underwent bacterial cultures on admission.
The material for bacterial culture was a pharyngeal swab.The hospital selected in the present study consisted of an internal medicine ward with long stay units(307 beds).The average duration of hospitalization was 545 days.We received about 40 patients from acute hospital units and about 15 patients were transferred to acute hospital units from our hospital every fiscal year.
In order to evaluate the various factors which may influence the occurrence of MRSA isolation, a cross sectional study was carried out.Some factors which may influence MRSA isolation were investigated, including medical and residential status prior to the admission such as own home, nursing home or other hospital, the age and gender of patients, an evaluation of activities of daily living(ADL), the nutritional state as expressed by hemoglobin or serum albumin, fever and infectious condition expressed by white blood cell count or c-reactive protein(CRP).
The ADL score was determined in the same manner as in our previous studies r. 7,8,9,10,11,12).The ADL score was the sum of the ability to perform ADL.Those who could take meals by themselves received one point for dining function, those who could walk by themselves got one point for walking function and those who could urinate and have bowel movements by themselves were given one point for toilet function.The patients without limitation got three points, those with partially limited ADL were given one or two points, and those with completely limited ADL received zero points.The information on ADL was collected from medical records.
Diagnosis of MRSA was made by the microdilution in susceptibility test for oxacillinu).Staphylococcus aureus, which showed resistance to oxacillin (minimum inhibitory concentration 4 g/ml), was diagnosed as MRSA.Serum albumin was measured by using autoanalyzer (Olympus, Tokyo).Statistical analysis was performed using the Statistical Analysis System package (SAS institute) 14).Significance was determined by the one way analysis of variance and the chisquare test.A multiple logistic regression analysis was used to control for confounding factors for the isolation of MRSA.The odds ratio(OR) and their 95% confidence interval(95% CI) were then calculated for each factor on the basis of logistic regression coefficient and its standard error.

RESULTS
As shown in Table 1, compared with patients transferred from other hospitals, those admitted from their own homes had a significantly smaller isolation rate of MRSA(p<0.05).The proportion of MRSA among the isolated S. aureus was smallest in the patients from their own homes and greatest in those from other hospitals.
As shown in Table 2, the patients from nursing homes were older than those from their own homes(p<0.01)and those from other hospitals(p<0.05).
The proportion of males was greater in the patients from their own homes than in those from nursing homes(p<0.01)and those from other hospitals(p<0.05),and females were more commonly seen in those from nursing homes than those from other hospitals (p<0.01).Compared with the patients from nursing homes or other hospitals, ADL scores(p<0.01) and the proportion of the patients with the ability to perform toilet functions (p<0.01),dining functions (p<0.01) or walking functions(p<0.01) were greater in the patients from their own homes.Patients from their own homes had higher levels of hemoglobin or serum albumin than those from nursing homes or other hospitals(p<0.01).(Table 3).
As shown in the Table 4, those patients with hypoalbuminemia had lower ADL scores than those without.Table 5 shows the result of multiple regression analysis in relation to the isolation of MRSA.Fever( + vs -; OR=1.60, 95% CI =1.06 -2.34) remained to be the strongest risk factor and disability in performing ADL( ADL score 0-1 vs 2-3; OR=1.49, 95% CI=1.02-2.18)was a significant risk factor for the MRSA isolation (model 1).The OR of hypoalbuminemia was 1.33 but failed to show statistical significance.
Multiple logistic regression analysis in a model excluding ADL revealed that fever( OR=1.77, 95% CI=1.18-2.66)was the strongest risk factor for MRSA isolation as well(model 2).Hypoalbuminemia became a significant risk factor in this model (OR= 1.55,95% CI=1.04-2.37).

DISCUSSION
Lower risk was observed in those patients from their own homes; the isolation rate of MRSA was 3.0% for those admitted from their own homes, 9.7% for those from nursing homes and 14.0% for those from other hospitals.This result is similar to the result of our previous study reported in 1993 7), in which the positive rate of MRSA in the elderly with infection was 0.0% for outpatients, 5.0% for nursing home residents and 27.0% for inpatients.This result may be partly explained by the difference of ADL and nutritional status among the three groups used in the present study; ADL scores(p<0.01) as well as the proportion of patients with the ability to perform toilet functions(p<0.01),dining functions(p<0.01) or walking functions (p<0.01) was greater in patients from their own homes compared with patients from nursing homes or other hospitals.
In addition, the patients from their own homes had higher levels of hemoglobin or serum albumin than those from nursing homes or other hospitals (p<0.01).
In the present study, the multiple logistic regression analysis revealed that lower ADL score was the strong risk factor for the MRSA isolation.This result is consistent with our previous studies 7.8,10,11,12), demonstrating that the MRSA positive patients had greater limitations in ADL than did the negative patients, and also that the risk of MRSA increased according to the limitation of ADL even after controlling for the effects of other factors such as antibiotic use and hypoalbuminemia.
Colonized and infected patients represent the major reservoir of MRSA in hospitals or nursing homes, and MRSA is most frequently spread from patient to patient on the hands of health care workers15>.Therefore, MRSA infection may occur more often in patients with a limited ADL, who need frequent help in their daily lives and experience hard contact with others many times a day.Therefore, there must be appropriate facilities for staff hand washing16), and the quality of care should be inspected regularly by each district health authority as well as monitored at operational levelsla.In contrast, age did not increase the risk of the MRSA isolation.This result is consistent with our previous studies 7. 8, 9.10,11.12)which have demonstrated that old age was not a risk factor of MRSA infection in the elderly.It is true that the elderly population is one of the high risk groups of MRSA infection.However, the results of the present study and our previous studies 7. 8 ,9,10,11,12) suggest that ADL is a more important factor than the actual age among the elderly population.
As shown in Table 3, the level of serum albumin was highest in the patients from their own homes, whose MRSA isolation rate was the lowest.This result supports the result of our previous studies, which showed that the level of serum albumin was significantly lower in the MRSA positive patients than in the negative patients 9.11.12) or control subjects10).Salive et al. reported that hypoalbuminemia was associated with limitations in ADL among the elderly population18).
In the present study, as shown in the Table 4, the patients with hypoalbuminemia had lower ADL score than those without .Our results are consistent with their study18).
In the present study, multivariate analysis also revealed that fever(OR= 1.60, 95%CI=1.06-2.34)was a strong risk factor for the MRSA isolation.Since the use of antibiotics is a risk factor of MRSA infection 7.8,9,10,11), this result may be partly explained by the fact that those with fever had greater probability of receiving antibiotics.Watanabe et al19) suggested that a secondary bacterial infection occurs at an early stage of the disease (2-3 days after viral infection).Therefore, the prevention of fever such as influenza-like illness seems to be important to prevent MRSA infection.
In conclusion, the elderly in nursing homes or in hospitals are a high risk group for the MRSA isolation because they have poor nutritional status and limited ADL.In the rapidly aging society of Japan, a lot of the demented and/or the handicapped elderly live in nursing homes or the geriatric hospitals with long stay units.Closed or crowded situations have been reported to have a relation to the out-break of influenza-like ill-ness20) and a secondary infection can occur at an early stage of viral infectious disease19).Our previous study21)revealed that infectious disease such as pneumonia was the first major cause of death(47.8% of total death) for demented and the third major cause of death (21.4%) for non-demented elderly in a nursing home.In addition, the present study revealed that fever such as influenza-likeillness may be the strongest risk factor of MRSA isolation.These findings strongly convince us of the importance of preventing influenza-like illness among the institutionalized or hospitalized elderly.
Very recently, Hirota et a1.22) pointed out that in Japan a markedly decreased influenza vaccine distribution is occurring, while the prevention of influenza is a major public problem and a program of vaccination in the elderly has been promoted in the United States23).Geriatrists as well as public health specialists in Japan need to recognize the importance of infectious control in nursing homes and geriatric hospitals.

Table 1 .
The isolation rate of MRSA and its proportion of Staphylococcus aureus from the pharyngeal swab cultures on admission according to their residential status before admission such as their own homes, nursing homes and hospitals.

Table 2 .
Age, sex, fever and activities of daily livings on admission among the patients from their own homes, nursing homes and other hospitals.Values are means ±SD, or number(%).**p<0.01vsnursing homes, #;p<0.05,##;p<0.01vsother hospitals Fever; body temperature on admission is more than 37.51 ADL score; activities of daily livings score is the sum of the ability to perform ADL.Those who could take meals by themselves received one point for dining function, those who could walk by themselves got one point for walking function and those who could urinate and have bowel movement by themselves were given one point for rest room function.

Table 3 .
Laboratory findings on admission among the patients from their own homes, nursing homes and other hospitals,

Table 4 .
Comparison of ADL score between the patients with and without hypoalbuminemia.
Values are means ± SD, **;P<0.01ADLscore; activities of daily livings score is the sum of the ability to perform ADL.Those who could take meals by themselves received one point for dining function, those who could walk by themselves got one point for walking function and those who could urinate and have bowel movement by themselves were given one point for rest room function.

Table 5 .
Multiple logistic regression analysis in relation to the isolation of MRSA.All factors in the tables are included in the analysis in this model.Model*2; ADL score is not included in the analysis in this model.ND; This factor is not included in the analysis.