Acta med. scand. Vol. 199, pp. 151-155, 1976
Bacteriological and Clinical Evaluation of Different Dialysate Delivery Systems Steen Gamwell Dawids and Rene Vejlsgaard From Medical Department P.Division of Nephrology, Rigshospitalet, and the Department of Clinical Microbiology9 Institute of Medical Microbiology, University of Copenhagen, Copenhagen, Denmark
ABSTRACT. In the period 1-74 four different dialysate delivery systems have been used in our department. 1 ) Central mixing of dialysate using tap water and a dialysate delivery line with “dead ends” resulting in stagnant dialysate. 2) Central mixing of dialysate with cold distilled water, otherwise equal to system 1. 3) Local mixing of dialysate with cold distilled water, delivered through a line with “dead ends” resulting in stagnant water. 4) Local mixing of dialysate with distilled water, cooled to 25°C just prior to use, reduced “dead ends” and monitored constant overflow to drain through the water supply line. The bacterial contamination of the four systems was examined and related to the clinical occurrence of pyrogenic and other reactions. An improvement was noted with the change from central to local mixing of dialysate (system 3) but complete sterility and virtual freedom from clinical reactions were first obtained in system 4. It is concluded that the use of sterile or near sterile dialysate is recommendable.
dialysate, exceeding lo6 col./ml is associated with an increasing frequency of severe pyrogenic reactions (12). The reported frequency of all types of pyrogenic reactions (mild and severe) varies in different materials. Excluding febrile reactions associated with blood transfusions, demonstrable infections and drugs, Robinson and Rosen (14) found that about 10% of dialytic treatments result in elevation of body temperature, often accompanied by symptoms of varying severity (e.g. vomiting, abdominal or back pain and hypotension). Since 1964 four different types of dialysate delivery systems have consecutively been used in our department, systems which were designed to be of increasing value in the prevention of bacterial contamination of the dialysate. It is the purpose of the present paper to report on the degree of contamination and its correlation to the clinical symptoms in these four systems.
Although it is usually accepted that the dialysis membrane is relatively impermeable to bacteria (and perhaps also to vira), pyrogenic reactions are well known symptoms during dialysis treatment with a heavily contaminated dialysate (14). Pyrogenic reactions are mostly due to the passage of bacterial toxins through the dialysis membrane, although direct passage of bacteria can occur (9, 11, 12, 13, 15, 16). It has been demonstrated that antibodies to dialysable bacterial endotoxins are present in the serum of many dialysis patients (9, 11). Further it has been found, that a bacterial count in the
DESCRIPTION OF THE FOUR DIALYSATE DELIVERY SYSTEMS System I (Fig. 1) uses tap water, batch mixing of dialysate subsequently heated to 25°C. dialysate delivery line with stagnating dialysate and non-heat sterilizable monitors. Wafer was of cold deep subsoil quality (81 l T ) . Mixing of tap water with pyrogen-tested dry salts or with sterile salt concentrates was carried out in a stainless steel tank containing I000 1. Dialysate delivery line was made of pyrex glass with teflon fittings. It had “dead ends” not only at the terminal point, but also at each individual tap site. The dialysate was run from the mixing tank through the line into local tanks, each containing about 150 I (not shown in Fig. 1). Monitors contained a heating unit to increase dialysate temperature to Actu med. scund. 199
152
S . G . Dawids and R . Vejlsgaard TAP- WATER
CONTINUOUS
JEi I-
PLANT
T
Fig. 1. Diagram of system 1 using tap water and batch
mixing.
Fig. 3. Diagram of system 3 using distilled water via storage tank and local mixing.
38°C and an effluent centrifugal pump. The entire dialysate delivery system and the monitors were cleaned and sterilized by flushing with hot tap water after each use and filled with 2 % formalin solution. Prior to use the formalin was washed out with hot tap water. System 2 (Fig. 2 ) uses 25°C distilled water from a distillation plant which has been described in an earlier paper (3) but is otherwise identical to system 1. System 3 (Fig. 3) uses distilled water delivered at 25°C from a storage tank with continuous overflow. The delivery line for distilled water is a single pipe of stainless steel with stagnant water in the whole length when not in use. Local mixing of dialysate is performed in monitors, which are sterilizable by heat and chemicals. This equipment has earlier been described in detail (4, 5 ) . Cleaning and sterilization of the delivery line for distilled water as in systems 1 and 2. System 4 (Fig. 4) uses hot distilled, sterile water cooled just before use to 25°C. The water delivery line has reduced dead ends at the tap sites, and has a continuous throughflow of water, regardless whether the dialysis stations are in use or not (3). Local mixing of distilled water and concentrate in sterilizable monitors as in system 3.
BACTERIOLOGICAL METHODS Cultures were taken from four comparable sites (A, B, C, D) in each system as indicated in Figs. 1-4. One or two portions of 500 ml fluid was collected aseptically in sterile bottles. The sample was drawn through a sterile microporous filter (Millipore@,0.45 p). The filter was then transferred aseptically to the surface of bloodyeast-agar substrate in a Petri dish and incubated at 35°C for 24 and 48 hours. The number of colonies was counted with a magnifying-glass. Identification of the colonies was made by conventional technique.
REGISTRATION OF SYMPTOMS Pyrogenic reactions were defined as an increase in tem-
perature of at least 0.6T and a peak temperature of at least 37.7"C. Fever due to known active infections, to transfusions and to drugs was disregarded. Pyrogenic reactions were commonly associated with different symptoms of varying intensity and were graded accord-
CONTINUOUS
RESERVOIR PLANT
DISTILLATION
RESERVOIR
Fig. 2. Diagram of system 2 using distilled water and
Fig. 4. Diagram of system 4 using hot distilled water,
batch mixing.
cooled prior to use, and local mixing.
Acrn med. scnnd. 199
Dialysate delivery systems
Table I . Pattern of bacterial counts in systems 1-4
fever. Disregarding dysequilibrium syndrome, too rapid ultrafiltration and other conditions, not associated with the dialytic procedure, such symptoms may be due to endotoxins as recently shown by Gazenfeldt-Gazit and Ehahou (9). In the present material they were graded as follows: Slight symptoms: severe headache, episode of vomiting, transient abdominal pain, transient BP drop. Severe symptoms: repeated vomiting, low back or abdominal pain, marked BP drop.
Bacterial counts in sample (col ./I) Culture sites
No. of samples
0-50
(%)
51-500 (%)
'500 (%)
System I A
B C D System 2 A
B C D System 3 A
B C D System 4 A
B C D
160 397 3 93 428
100 87 73 76
0 4 6 7
0 9 21 17
I3 1 355 227 50 1
92 81 70 69
2 5 1
7
6 14 29 24
160
94 47 55 58
6 24 24 23
29 21 19
100
0 0 0 6
0 0 0 0
353 35 1 346 42 83 I33 133
I00 100 94
153
RESULTS
ingly: Grade I) NO associated symptoms. Grade 2) Slight symptoms (i.e. headache and/or vomiting and/or chills). Grade 3) Severe symptoms (i.e. BP drop and/or recurrent vomiting and/or pain in lower back or abdomen and/or pronounced malaise). Other reactions. Symptoms of grades 1-3 according to the definitions above could also occur in the absence of
Table I illustrates the percentage of consecutive samples which were found contaminated at the points A, B, C and D in the four different systems shown in Figs. 1 4 It appears that the degree of contamination increases from point A to D without any significant difference between systems 1-3, whereas system 4 is sterile for all practical purposes. Repeated testing of system 4 at irregular intervals over an extended period confirms this observation. It should be emphasized in this context that systems 1-3 remained infected despite the extensive, daily cleaning and sterilization procedures mentioned in the description Of the four dialYsate delivery systems. In contradistinction, system 4 remained sterile over an observation period of 4 years without any cleaning and sterilization of the water line. Table 11 shows the clinical symptoms graded
Table 11. Frequency of fever in the different systems and relation to accompanying symptoms Other reactions (%)
Pyrogenic reactions (%) System
symptoms
Slight symptoms
Severe symptoms
Total
No symptoms
Slight symptoms
Severe symptoms
I 2 3 4
2.2 4.2 I .6 1.5
5.7 2.2 2.2 0.3
4.5 0.8 1.6 0.0
12.4 7.8 5.8 1.8
59.6 55.4 79.6 95.6
17.3 33.3 9.7 2.5
10.6 4. I 5.3 0.1
No
Table 111. Bacterial species related to symptoms Symptoms (%)
Ps. aeruginosa Ps. stutzeri Enterobact. cloacae Klebsiella oxytoca Bacterium anitratum
No. of samples
No
Slight
Severe
Comments
194 I15 30 24 8
32 12 14 0 0
63 51 71 42 37
5 37 15 58 63
Mainly headache and fever BP instability, vomiting, malaise Headache and BP instability Headache and BP instability BP instability Acta med. wand. 199
S . G . Dawids and R . Vejlsgaard
154
100%
-
50 %
-
0% . .~
-
SYSTEM
z z z
z z z
z z z
N o symptoms
Slight sympioms
Severe symptoms
1 2 3
1 2 3
1 2 3
0
0-50
germs/\
51 -500 germs / I
5
to dominate other bacteria once it is established in the pipe system. The situation becomes dangerous once a change in the species occurs (2). Especially Klebsiella oxytoca and B. anitratum appear to give rise to frequent reactions. The main clinical finding in connection with these bacterial contaminants is pronounced instability of BP and circulatory failure. Despite almost sterile conditions a low frequency of fever is seen in system 4, obviously of non-bacterial origin. The results from 7 out of 27 patients in a prospective study over 3 months is shown in Table IV. The 7 patients all have in common pronounced symptoms of anemia necessitating regular transfusions. Only one of the remaining 20 patients received regular transfusions (lO/year).
> 500 germs / I
Fig. 5. Correlation between symptoms and bacterial contamination of the dialysate in systems 1, 2 and 3.
according to their severity. Although the contamination of systems 1-3 is of the same order of magnitude, it appears that system 3 (with local mixing of water and concentrate) is better than systems 1 and 2 with central mixing. However, only system 4 shows a very low incidence of pyrogenic and other reactions. Fig. 5 shows the relation between severity of symptoms (irrespective of the presence of fever) in relation to the degree of bacterial contamination. There seems to be a clear-cut correlation between the seventy of symptoms and the degree of contamination in systems 1, 2 and 3. The relation between the species of bacteria and the clinical symptoms has been evaluated. In systems 1 and 2 more than 70 % of the significantly contaminated samples (>SO0 col./l) contained two or more species which makes it difficult to evaluate the reactivity to any specific bacteria. Analysis of the material does, however, indicate that a mixed contamination gives rise to a higher frequency and severity of reactions. In order to evaluate possible differences in the clinical reactivity to specific bacteria, 371 dialysate samples which were contaminated with only one fully identified bacterial species were considered. The results are shown in Table 111. Ps. aemginosa seems quite harmless from a clinical point of view, which is fortunate as this species is the most commonly found and tends Acla med. scand. 199
DISCUSSION Despite the introduction of sterile, disposable dialysers clinical reactions caused by bacterial contamination of the dialysate are still in evidence. The frequency has been reported to be about 10% (14) and a good correlation has been found between the degree of contamination and the severity of the reactions (6,7, 9, 10, 12). The latter finding is supported by the present study as shown in Fig. 5. The present study also permits an evaluation of the correlation between the species of contaminating bacteria and clinical reactions. The results shown in Table 111 demon-
Table IV. Pyrogenic reactions during 752 hemodialyses Total
Pat. no.
Fever during dialysis No symptoms Slight symptoms Severe symptoms Approx. transfusionslv. Cultures'iaken Growth in water Growthinsystem Growthin blood
1
2
3 4
5
6 7
n
2 2 0 0
6
3 3 0 0
2 0 2
1
172.3 14 1.9 3 0.4
5 1
0
2 2 0 0
0
1
I 1 0 0 0 0
8 24 12 8 12 18 4 2 6 3 2 2 I I 0 0 0 0 0 0 0 1 0 0 0 (I) 0 0 0 0 0 0 0 0 0
%
0
17 0 2 0
Treated topically in shunt with thrombolytic enzymes (Brinase").
Dicilysate delivery systems
strate that contamination of the dialysate with Klebsiella oxytoca and Bacterium anitratum is more hazardous than the three other commonly occurring contaminants in our systems. Off-hand, one might expect that local mixing of the dialysate and the use of fully sterilizable monitors might eliminate the problem. A reduction in the frequency and in the severity of reactions was in fact observed when these precautions were introduced (cf. system 3 to systems I and 2 in Table 111). It is quite clear, however, that the precautions are not sufficient per se. A dramatic reduction in the number of clinical reactions was first observed when the entire system was kept sterile throughout (system 4, Table 111). The very few reactions, still occurring in this situation, must apparently be explained by factors unrelated to the quality of the dialysate. System 4 has previously been described in detail (3). In brief, it operates with freshly produced sterile (and distilled) water cooled shortly before use and delivered through a pipeline with continous overflow, avoiding all possibilities for stagnation. The water is mixed locally with concentrate in monitors which are fully sterilizable (by heat and chemicals) between each use. The expense of this system has previously been calculated to be 0.032 D. kr./l water (3). However, since the pipeline remains sterile automatically, personnel hours for cleaning and sterilization of the pipeline are saved. In conclusion we find that the use of dialysate with the lowest possible degree of contamination is highly recommendable. In a center this may be achieved by introducing a system which has allor at least most of-the merits of our system 4. In the home such a system is hardly economically acceptable, but it is recommendable to take all precautions to prevent contamination of those compounds which are most liable to permit bacterial multiplication.
155
REFERENCES 1. Cole, J. J., Fritzen, J. R., Vizzo. J . E., van Paas-
schen, W. H. & Grimsrud, L.: One year's experience with a central dialysate supply system in a hospital. Trans. Amer. SOC.artifintern. Org. I1:22, 1965. 2. Curtis, J. R., Wing, A. J. & Coleman, J. C.: Bacillus cereus bacteraemia. A complication of intermittent haemodialysis. Lancet I: 136, 1967. 3. Dawids, S. G.: Large scale production of sterile, distilled water for hospital dialysis. Acta med. scand. 193:387, 1973. 4. Dawids, S . G., Boe, C. & Andreasen, M. M.: A new hemodialysis console. I. General description. Acta med. scand. 193: 373, 1973. 5. - A new hemodialysis console. 11. Description of components. Acta med. scand. 193: 379, 1973. 6. Favero, M. S . , Carson, L. A., Bond, W. W. & Petersen, N. J.: Factors that influence microbial contarnination of fluids associated with hemodialysis machines. Appl. Microbiol. 2: 822, 1974. 7. Favero, M. S., Petersen, N. J., Boyer, K. M., Carson, L. A. & Bond, W. W.: Microbial contamination of renal dialysis systems and associated health risks. Trans. Amer. SOC.artif. intern. Org. XX: 175, 1974. 8. Fregerslev, S.: Microbial content in a non-sterile concentrated hemodialysis solution. Dansk T. Farm. 44: 332, 1970. 9. Gazenfeldt-Gazit, E. & Ehahou. H. E.: Endotoxin antibodies in patients on maintenance hemodialysis. Israel J. med. Sci. 5: 1032, 1969. 10. Jones, D. M., Tobin, B. M., Harlow, G. R. & Ralston, A. J.: Bacteriological studies of the modified Kiil dialyzer. Brit. med. J. 3: 135, 1970. 11. - Antibody production in patients on regular hemodialysis to organisms present in dialysate. Proc. Europ. Dial. Trans. Ass. 9: 575, 1972. 12. Kidd, E. E.: Bacterial contamination of dialyzing fluid of artificial kidney. Brit. med. J. 1:880, 1964. 13. Quarles, J. M., Belding, R. C., Beaman, T. C. & Gerhardt, P.: Hemodialysis culture of Serratia marescens in a goat-artificial kidney-fermentor system. Infect. Immunity 9: 550, 1974. 14. Robinson, P. J. A. & Rosen, S . M.: Pyrexial reactions during hemodialysis. Brit. med. J. 1:528, 1971. 15. Tierno, P. M. & Aboody, R.: Risk of bacterial infection resulting from a blood leak during hemodialysis. Nephron 6: 110, 1%9. 16. US Government Printing Ofice, Washington D.C.: Hemodialysis manual DHEN (HSM) 72-7002, 1971.
Acra med. scand. 199
Bacteriological and Clinical Evaluation of Different Dialysate Delivery Systems Steen Gamwell Dawids and Rene Vejlsgaard From Medical Department P.Division of Nephrology, Rigshospitalet, and the Department of Clinical Microbiology9 Institute of Medical Microbiology, University of Copenhagen, Copenhagen, Denmark
ABSTRACT. In the period 1-74 four different dialysate delivery systems have been used in our department. 1 ) Central mixing of dialysate using tap water and a dialysate delivery line with “dead ends” resulting in stagnant dialysate. 2) Central mixing of dialysate with cold distilled water, otherwise equal to system 1. 3) Local mixing of dialysate with cold distilled water, delivered through a line with “dead ends” resulting in stagnant water. 4) Local mixing of dialysate with distilled water, cooled to 25°C just prior to use, reduced “dead ends” and monitored constant overflow to drain through the water supply line. The bacterial contamination of the four systems was examined and related to the clinical occurrence of pyrogenic and other reactions. An improvement was noted with the change from central to local mixing of dialysate (system 3) but complete sterility and virtual freedom from clinical reactions were first obtained in system 4. It is concluded that the use of sterile or near sterile dialysate is recommendable.
dialysate, exceeding lo6 col./ml is associated with an increasing frequency of severe pyrogenic reactions (12). The reported frequency of all types of pyrogenic reactions (mild and severe) varies in different materials. Excluding febrile reactions associated with blood transfusions, demonstrable infections and drugs, Robinson and Rosen (14) found that about 10% of dialytic treatments result in elevation of body temperature, often accompanied by symptoms of varying severity (e.g. vomiting, abdominal or back pain and hypotension). Since 1964 four different types of dialysate delivery systems have consecutively been used in our department, systems which were designed to be of increasing value in the prevention of bacterial contamination of the dialysate. It is the purpose of the present paper to report on the degree of contamination and its correlation to the clinical symptoms in these four systems.
Although it is usually accepted that the dialysis membrane is relatively impermeable to bacteria (and perhaps also to vira), pyrogenic reactions are well known symptoms during dialysis treatment with a heavily contaminated dialysate (14). Pyrogenic reactions are mostly due to the passage of bacterial toxins through the dialysis membrane, although direct passage of bacteria can occur (9, 11, 12, 13, 15, 16). It has been demonstrated that antibodies to dialysable bacterial endotoxins are present in the serum of many dialysis patients (9, 11). Further it has been found, that a bacterial count in the
DESCRIPTION OF THE FOUR DIALYSATE DELIVERY SYSTEMS System I (Fig. 1) uses tap water, batch mixing of dialysate subsequently heated to 25°C. dialysate delivery line with stagnating dialysate and non-heat sterilizable monitors. Wafer was of cold deep subsoil quality (81 l T ) . Mixing of tap water with pyrogen-tested dry salts or with sterile salt concentrates was carried out in a stainless steel tank containing I000 1. Dialysate delivery line was made of pyrex glass with teflon fittings. It had “dead ends” not only at the terminal point, but also at each individual tap site. The dialysate was run from the mixing tank through the line into local tanks, each containing about 150 I (not shown in Fig. 1). Monitors contained a heating unit to increase dialysate temperature to Actu med. scund. 199
152
S . G . Dawids and R . Vejlsgaard TAP- WATER
CONTINUOUS
JEi I-
PLANT
T
Fig. 1. Diagram of system 1 using tap water and batch
mixing.
Fig. 3. Diagram of system 3 using distilled water via storage tank and local mixing.
38°C and an effluent centrifugal pump. The entire dialysate delivery system and the monitors were cleaned and sterilized by flushing with hot tap water after each use and filled with 2 % formalin solution. Prior to use the formalin was washed out with hot tap water. System 2 (Fig. 2 ) uses 25°C distilled water from a distillation plant which has been described in an earlier paper (3) but is otherwise identical to system 1. System 3 (Fig. 3) uses distilled water delivered at 25°C from a storage tank with continuous overflow. The delivery line for distilled water is a single pipe of stainless steel with stagnant water in the whole length when not in use. Local mixing of dialysate is performed in monitors, which are sterilizable by heat and chemicals. This equipment has earlier been described in detail (4, 5 ) . Cleaning and sterilization of the delivery line for distilled water as in systems 1 and 2. System 4 (Fig. 4) uses hot distilled, sterile water cooled just before use to 25°C. The water delivery line has reduced dead ends at the tap sites, and has a continuous throughflow of water, regardless whether the dialysis stations are in use or not (3). Local mixing of distilled water and concentrate in sterilizable monitors as in system 3.
BACTERIOLOGICAL METHODS Cultures were taken from four comparable sites (A, B, C, D) in each system as indicated in Figs. 1-4. One or two portions of 500 ml fluid was collected aseptically in sterile bottles. The sample was drawn through a sterile microporous filter (Millipore@,0.45 p). The filter was then transferred aseptically to the surface of bloodyeast-agar substrate in a Petri dish and incubated at 35°C for 24 and 48 hours. The number of colonies was counted with a magnifying-glass. Identification of the colonies was made by conventional technique.
REGISTRATION OF SYMPTOMS Pyrogenic reactions were defined as an increase in tem-
perature of at least 0.6T and a peak temperature of at least 37.7"C. Fever due to known active infections, to transfusions and to drugs was disregarded. Pyrogenic reactions were commonly associated with different symptoms of varying intensity and were graded accord-
CONTINUOUS
RESERVOIR PLANT
DISTILLATION
RESERVOIR
Fig. 2. Diagram of system 2 using distilled water and
Fig. 4. Diagram of system 4 using hot distilled water,
batch mixing.
cooled prior to use, and local mixing.
Acrn med. scnnd. 199
Dialysate delivery systems
Table I . Pattern of bacterial counts in systems 1-4
fever. Disregarding dysequilibrium syndrome, too rapid ultrafiltration and other conditions, not associated with the dialytic procedure, such symptoms may be due to endotoxins as recently shown by Gazenfeldt-Gazit and Ehahou (9). In the present material they were graded as follows: Slight symptoms: severe headache, episode of vomiting, transient abdominal pain, transient BP drop. Severe symptoms: repeated vomiting, low back or abdominal pain, marked BP drop.
Bacterial counts in sample (col ./I) Culture sites
No. of samples
0-50
(%)
51-500 (%)
'500 (%)
System I A
B C D System 2 A
B C D System 3 A
B C D System 4 A
B C D
160 397 3 93 428
100 87 73 76
0 4 6 7
0 9 21 17
I3 1 355 227 50 1
92 81 70 69
2 5 1
7
6 14 29 24
160
94 47 55 58
6 24 24 23
29 21 19
100
0 0 0 6
0 0 0 0
353 35 1 346 42 83 I33 133
I00 100 94
153
RESULTS
ingly: Grade I) NO associated symptoms. Grade 2) Slight symptoms (i.e. headache and/or vomiting and/or chills). Grade 3) Severe symptoms (i.e. BP drop and/or recurrent vomiting and/or pain in lower back or abdomen and/or pronounced malaise). Other reactions. Symptoms of grades 1-3 according to the definitions above could also occur in the absence of
Table I illustrates the percentage of consecutive samples which were found contaminated at the points A, B, C and D in the four different systems shown in Figs. 1 4 It appears that the degree of contamination increases from point A to D without any significant difference between systems 1-3, whereas system 4 is sterile for all practical purposes. Repeated testing of system 4 at irregular intervals over an extended period confirms this observation. It should be emphasized in this context that systems 1-3 remained infected despite the extensive, daily cleaning and sterilization procedures mentioned in the description Of the four dialYsate delivery systems. In contradistinction, system 4 remained sterile over an observation period of 4 years without any cleaning and sterilization of the water line. Table 11 shows the clinical symptoms graded
Table 11. Frequency of fever in the different systems and relation to accompanying symptoms Other reactions (%)
Pyrogenic reactions (%) System
symptoms
Slight symptoms
Severe symptoms
Total
No symptoms
Slight symptoms
Severe symptoms
I 2 3 4
2.2 4.2 I .6 1.5
5.7 2.2 2.2 0.3
4.5 0.8 1.6 0.0
12.4 7.8 5.8 1.8
59.6 55.4 79.6 95.6
17.3 33.3 9.7 2.5
10.6 4. I 5.3 0.1
No
Table 111. Bacterial species related to symptoms Symptoms (%)
Ps. aeruginosa Ps. stutzeri Enterobact. cloacae Klebsiella oxytoca Bacterium anitratum
No. of samples
No
Slight
Severe
Comments
194 I15 30 24 8
32 12 14 0 0
63 51 71 42 37
5 37 15 58 63
Mainly headache and fever BP instability, vomiting, malaise Headache and BP instability Headache and BP instability BP instability Acta med. wand. 199
S . G . Dawids and R . Vejlsgaard
154
100%
-
50 %
-
0% . .~
-
SYSTEM
z z z
z z z
z z z
N o symptoms
Slight sympioms
Severe symptoms
1 2 3
1 2 3
1 2 3
0
0-50
germs/\
51 -500 germs / I
5
to dominate other bacteria once it is established in the pipe system. The situation becomes dangerous once a change in the species occurs (2). Especially Klebsiella oxytoca and B. anitratum appear to give rise to frequent reactions. The main clinical finding in connection with these bacterial contaminants is pronounced instability of BP and circulatory failure. Despite almost sterile conditions a low frequency of fever is seen in system 4, obviously of non-bacterial origin. The results from 7 out of 27 patients in a prospective study over 3 months is shown in Table IV. The 7 patients all have in common pronounced symptoms of anemia necessitating regular transfusions. Only one of the remaining 20 patients received regular transfusions (lO/year).
> 500 germs / I
Fig. 5. Correlation between symptoms and bacterial contamination of the dialysate in systems 1, 2 and 3.
according to their severity. Although the contamination of systems 1-3 is of the same order of magnitude, it appears that system 3 (with local mixing of water and concentrate) is better than systems 1 and 2 with central mixing. However, only system 4 shows a very low incidence of pyrogenic and other reactions. Fig. 5 shows the relation between severity of symptoms (irrespective of the presence of fever) in relation to the degree of bacterial contamination. There seems to be a clear-cut correlation between the seventy of symptoms and the degree of contamination in systems 1, 2 and 3. The relation between the species of bacteria and the clinical symptoms has been evaluated. In systems 1 and 2 more than 70 % of the significantly contaminated samples (>SO0 col./l) contained two or more species which makes it difficult to evaluate the reactivity to any specific bacteria. Analysis of the material does, however, indicate that a mixed contamination gives rise to a higher frequency and severity of reactions. In order to evaluate possible differences in the clinical reactivity to specific bacteria, 371 dialysate samples which were contaminated with only one fully identified bacterial species were considered. The results are shown in Table 111. Ps. aemginosa seems quite harmless from a clinical point of view, which is fortunate as this species is the most commonly found and tends Acla med. scand. 199
DISCUSSION Despite the introduction of sterile, disposable dialysers clinical reactions caused by bacterial contamination of the dialysate are still in evidence. The frequency has been reported to be about 10% (14) and a good correlation has been found between the degree of contamination and the severity of the reactions (6,7, 9, 10, 12). The latter finding is supported by the present study as shown in Fig. 5. The present study also permits an evaluation of the correlation between the species of contaminating bacteria and clinical reactions. The results shown in Table 111 demon-
Table IV. Pyrogenic reactions during 752 hemodialyses Total
Pat. no.
Fever during dialysis No symptoms Slight symptoms Severe symptoms Approx. transfusionslv. Cultures'iaken Growth in water Growthinsystem Growthin blood
1
2
3 4
5
6 7
n
2 2 0 0
6
3 3 0 0
2 0 2
1
172.3 14 1.9 3 0.4
5 1
0
2 2 0 0
0
1
I 1 0 0 0 0
8 24 12 8 12 18 4 2 6 3 2 2 I I 0 0 0 0 0 0 0 1 0 0 0 (I) 0 0 0 0 0 0 0 0 0
%
0
17 0 2 0
Treated topically in shunt with thrombolytic enzymes (Brinase").
Dicilysate delivery systems
strate that contamination of the dialysate with Klebsiella oxytoca and Bacterium anitratum is more hazardous than the three other commonly occurring contaminants in our systems. Off-hand, one might expect that local mixing of the dialysate and the use of fully sterilizable monitors might eliminate the problem. A reduction in the frequency and in the severity of reactions was in fact observed when these precautions were introduced (cf. system 3 to systems I and 2 in Table 111). It is quite clear, however, that the precautions are not sufficient per se. A dramatic reduction in the number of clinical reactions was first observed when the entire system was kept sterile throughout (system 4, Table 111). The very few reactions, still occurring in this situation, must apparently be explained by factors unrelated to the quality of the dialysate. System 4 has previously been described in detail (3). In brief, it operates with freshly produced sterile (and distilled) water cooled shortly before use and delivered through a pipeline with continous overflow, avoiding all possibilities for stagnation. The water is mixed locally with concentrate in monitors which are fully sterilizable (by heat and chemicals) between each use. The expense of this system has previously been calculated to be 0.032 D. kr./l water (3). However, since the pipeline remains sterile automatically, personnel hours for cleaning and sterilization of the pipeline are saved. In conclusion we find that the use of dialysate with the lowest possible degree of contamination is highly recommendable. In a center this may be achieved by introducing a system which has allor at least most of-the merits of our system 4. In the home such a system is hardly economically acceptable, but it is recommendable to take all precautions to prevent contamination of those compounds which are most liable to permit bacterial multiplication.
155
REFERENCES 1. Cole, J. J., Fritzen, J. R., Vizzo. J . E., van Paas-
schen, W. H. & Grimsrud, L.: One year's experience with a central dialysate supply system in a hospital. Trans. Amer. SOC.artifintern. Org. I1:22, 1965. 2. Curtis, J. R., Wing, A. J. & Coleman, J. C.: Bacillus cereus bacteraemia. A complication of intermittent haemodialysis. Lancet I: 136, 1967. 3. Dawids, S. G.: Large scale production of sterile, distilled water for hospital dialysis. Acta med. scand. 193:387, 1973. 4. Dawids, S . G., Boe, C. & Andreasen, M. M.: A new hemodialysis console. I. General description. Acta med. scand. 193: 373, 1973. 5. - A new hemodialysis console. 11. Description of components. Acta med. scand. 193: 379, 1973. 6. Favero, M. S . , Carson, L. A., Bond, W. W. & Petersen, N. J.: Factors that influence microbial contarnination of fluids associated with hemodialysis machines. Appl. Microbiol. 2: 822, 1974. 7. Favero, M. S., Petersen, N. J., Boyer, K. M., Carson, L. A. & Bond, W. W.: Microbial contamination of renal dialysis systems and associated health risks. Trans. Amer. SOC.artif. intern. Org. XX: 175, 1974. 8. Fregerslev, S.: Microbial content in a non-sterile concentrated hemodialysis solution. Dansk T. Farm. 44: 332, 1970. 9. Gazenfeldt-Gazit, E. & Ehahou. H. E.: Endotoxin antibodies in patients on maintenance hemodialysis. Israel J. med. Sci. 5: 1032, 1969. 10. Jones, D. M., Tobin, B. M., Harlow, G. R. & Ralston, A. J.: Bacteriological studies of the modified Kiil dialyzer. Brit. med. J. 3: 135, 1970. 11. - Antibody production in patients on regular hemodialysis to organisms present in dialysate. Proc. Europ. Dial. Trans. Ass. 9: 575, 1972. 12. Kidd, E. E.: Bacterial contamination of dialyzing fluid of artificial kidney. Brit. med. J. 1:880, 1964. 13. Quarles, J. M., Belding, R. C., Beaman, T. C. & Gerhardt, P.: Hemodialysis culture of Serratia marescens in a goat-artificial kidney-fermentor system. Infect. Immunity 9: 550, 1974. 14. Robinson, P. J. A. & Rosen, S . M.: Pyrexial reactions during hemodialysis. Brit. med. J. 1:528, 1971. 15. Tierno, P. M. & Aboody, R.: Risk of bacterial infection resulting from a blood leak during hemodialysis. Nephron 6: 110, 1%9. 16. US Government Printing Ofice, Washington D.C.: Hemodialysis manual DHEN (HSM) 72-7002, 1971.
Acra med. scand. 199
