Among the 294 patients on HD therapy in both centers at the beginning of the data collection, 160 participants were studied/analyzed. (Figure 1)
Baseline Socio-demographic-clinical characteristics of participants
As planned, 160 patients were included in the evaluation. 17.1% of the males and 32.7% of the females resulted in moderate/severe MICS status. The mean age (±SD) of patients was 52.2(13.3) years and the mean duration on HD therapy was 22 (+18) months (range: 3 to 126 months). Most of them, 132 (82.5%) were receiving three times per week HD therapy and 76 (47.5%) had an arteriovenous fistula (AVF) as the vascular access in use for HD. The majority of them, 136 (85%) were living with a partner, and 144 (90%) were covered by health insurance. Comorbidity-wise, almost all 154 (96.3%) had hypertension that was as well the commonest (43.1%) reported cause of End-Stage Renal Disease (ESRD). Diabetes mellitus was prevalent in 43.1% and anemia was present in 153 (95.6 %) patients. (Table 1)
Table 1. Socio-demographic and clinical characteristics of patients in relation to severity of MICS (N=160)
Characteristics
|
n (%)
|
MICS status
|
p-value
|
No
|
Mild
|
Mod-Severe
|
Age groups (years)
18-39
40 – 59
> 60
|
30 (18.8%)
75 (46.9%)
55 (34.4%)
|
12 (40.0%)
46 (61.3%)
28 (50.9%)
|
9 (30.0%)
17 (22.7%)
13 (23.6%)
|
9 (30.0%)
12 (16.0%)
14 (25.5%)
|
0.310a
|
Mean Age (+ SD) years
|
52.2 + 13.3
|
53.2 + 11.8
|
49.7 + 14.4
|
52.6 + 15.3
|
0.397c
|
Gender
Male
Female
|
111 (69.4%)
49 (30.6%)
|
65 (58.6%)
21 (42.9%)
|
27 (24.3%)
12 (24.5%)
|
19 (17.1%)
16 (32.7%)
|
0.072a
|
Marital Status
Living with a partner
Single
|
136 (85.0%)
24 (15.0%)
|
74 (54.4%)
12 (50.0%)
|
32 (23.5%)
7 (29.2%)
|
30 (22.1%)
5 (20.8%)
|
0.837a
|
Level of Education
No Formal / Primary
Post-primary
|
41 (25.6%)
119 (74.4%)
|
19 (46.3%)
67 (56.3%)
|
13 (31.7%)
26 (21.8%)
|
9 (22.0%)
26 (21.8%)
|
0.429a
|
Mode of Payment
Health Insurance
Out Of Pocket
|
144 (90.0%)
16 (10.0%)
|
78 (54.2%)
8 (50.0%)
|
34 (23.6%)
5 (31.2%)
|
32 (22.2%)
3 (18.8%)
|
0.776b
|
Frequency of HD
Thrice/week
Twice/week
|
132 (82.5%)
28 (17.5%)
|
73 (55.3%)
13 (46.4%)
|
30 (22.7%)
9 (32.1%)
|
29 (22.0%)
6 (21.4%)
|
0.575a
|
Duration of HD (years)
< 1
1-4
> 4
|
54 (33.8%)
92 (57.4%)
14 (8.8%)
|
34 (63.0%)
49 (53.3%)
3 (21.4%)
|
16 (29.6%)
22 (23.9%)
1 (7.1%)
|
4 (7.4%)
21 (22.8%)
10 (71.4%)
|
<0.001b
|
Duration on HD (months)
Mean (+ SD)
|
22 + 18
|
17.9 + 12.3
|
18 + 13.5
|
34.7 + 26.3
|
<0.001c
|
Vascular Access
AV Fistula
Central Venous Catheter
|
76 (47.5%)
84 (52.5%)
|
42 (55.3%)
44 (52.4%)
|
16 (21.1%)
23 (27.4%)
|
18 (23.7%)
17 (20.2%)
|
0.644a
|
Dialysis Adequacy
Inadequate (URR < 65%)
Adequate (URR > 65%)
|
38 (23.8%)
122 (76.3%)
|
19 (50.0%)
67 (54.9%)
|
8 (21.1%)
31 (25.4%)
|
11 (28.9%)
24 (19.7%)
|
0.479a
|
Diabetes Mellitus
Yes
No
|
69 (43.1%)
91 (56.9%)
|
40 (58.0%)
46 (50.5%)
|
12 (17.4%)
27 (29.7%)
|
17 (24.6%)
18 (19.8%)
|
0.198a
|
Hypertension
Yes
No
|
154 (96.3%)
6 (3.7%)
|
83 (53.9%)
3 (50.0%)
|
37 (24.0%)
2 (33.3%)
|
34 (22.1%)
1 (16.7%)
|
0.863b
|
Hypercholesterolaemia
Yes (TC > 220mg/dl)
No (TC < 220mg/dl)
|
9 (5.6%)
151 (94.4%)
|
8 (88.9%)
78 (51.7%)
|
1 (11.1%)
38 (25.2%)
|
0 (0.0%)
35 (23.2%)
|
0.133b
|
HIV Infection
Yes
No
|
15 (9.4%)
145 (90.6%)
|
5 (33.3%)
81 (55.9%)
|
5 (33.3%)
34 (23.4%)
|
5 (33.3%)
30 (20.7%)
|
0.223b
|
HBV Infection
Yes
No
|
9 (5.6%)
151 (94.4%)
|
5 (55.6%)
81 (53.6%)
|
3 (33.3%)
36 (23.8%)
|
1 (11.1%)
34 (22.5%)
|
0.740b
|
a Chi-square test, b Fisher’s exact test, c ANOVA
Prevalence and Severity of Malnutrition Inflammation Complex Syndrome (MICS)
Of the 160 patients on maintenance haemodialysis, 74 (46.3%) had MICS and these were categorized as mild MICS in 39 (24.4%) and moderate to severe MICS in 35 (21.9%) of the patients. (Figure 2)
Malnutrition Inflammation Score and its parameters
The Malnutrition-Inflammation-Score (MIS) ranged from 0 to 28 with a median MIS of 5 and a mean (±SD) MIS of 7.6 (±5.1). Compared to well-nourished patients, the mean MIS was significantly higher among underweight patients (14.8 vs 6.2, p<0.001). Overall dry weight loss >0.5kg over the preceding 3 months was present in 104 (65%) patients and severe dry weight loss (>5%) was present in only 13 (8.1%) patients. Dietary intake was good in almost half; 87 (54.4%) patients, and severe dietary limitations were present in only four (2.5%) patients. About one-third of the patients, 50 (31.2%) had some gastrointestinal symptoms. Physical examination revealed 98 (61.3%) patients had normal fat stores and 83 (51.9%) patients had no muscle wasting. Severe muscle wasting was present in 14 (8.8%) patients.
Body size, Inflammation and its association with MICS
Among the 160 patients, 27 (16.9%) were underweight and among the 103 patients whose CRP was measured, 84 (81.6%) had inflammation (CRP>5 mg/l). The presence of MICS was higher in underweight patients. Also, inflammation was significantly associated with MICS. (Table 2)
Table 2. Association between Body size and Inflammation with MICS
Body size (N=160)
|
MICS Present
|
MICS Absent
|
TOTAL
|
p-value
|
Underweight
|
26 (96.3%)
|
1 (3.7%)
|
27 (100%)
|
<0.001*
|
Normal weight
|
46 (48.4%)
|
49 (51.6%)
|
95 (100%)
|
Over-weight/Obese
|
2 (5.3%)
|
36 (94.7%)
|
38 (100%)
|
Inflammation (N=103)
|
Present
|
53 (63.1%)
|
31 (36.9%)
|
84 (100%)
|
<0.001*
|
Absent
|
2 (10.5%)
|
17 (89.5%)
|
19 (100%)
|
* Test Statistic: Fisher’s exact test
Factors associated with Malnutrition Inflammation Complex Syndrome
In univariate analysis female gender, longer duration on HD, hypercholesterolaemia, and having HIV associated with the occurrence of MICS. For multivariate analysis, the purposive selection of variables showing a univariate p<0.2 were included in the final regression logistic model. Patients with high cholesterol levels seemed to be protected against the development of MICS. Finally, patients longer on haemodialysis significantly associated with a higher risk for MICS. (Table 3)
Table 3. Factors associated with Malnutrition Inflammation Complex Syndrome (N=160)
Characteristics
|
CORa (95% CI)
|
p-value
|
AORb (95% CI)
|
p-value
|
Age groups
< 60 years
> 60 years
|
Ref
1.19 (0.62-2.3)
|
0.60
|
-
|
-
|
Gender
Male
Female
|
Ref
1.88 (0.95-3.72)
|
0.07
|
Ref
2.03 (0.97-4.25)
|
0.061
|
Marital Status
Living with a partner
Single
|
Ref
1.19 (0.5-2.84)
|
0.69
|
-
|
-
|
Mode of payment
Health Insurance
Paying out of pocket
|
Ref
1.18 (0.42-3.32)
|
0.75
|
-
|
-
|
Duration on HD
< 4 years
> 4 years
|
Ref
4.83 (1.29-18)
|
0.02
|
Ref
5.04 (1.33-19.2)
|
0.018
|
Frequency of HD
Thrice/week
Twice/week
|
Ref
1.43 (0.63-3.24)
|
0.39
|
-
|
-
|
Vascular Access
AV Fistula
Central Venous Catheter
|
Ref
1.12 (0.6-2.09)
|
0.72
|
-
|
-
|
Dialysis Adequacy
Inadequate (URR < 65%)
Adequate (URR > 65%)
|
1.22 (0.59-2.53)
Ref
|
0.60
|
-
|
-
|
Diabetes Mellitus
Yes
No
|
0.74 (0.4-1.4)
Ref
|
0.35
|
-
|
-
|
Hypertension
Yes
No
|
0.86 (0.17-4.37)
Ref
|
0.85
|
-
|
-
|
Hypercholesterolaemia
Yes (TC > 220mg/dl)
No (TC < 220mg/dl)
|
0.13 (0.02-1.1)
Ref
|
0.06
|
0.11 (0.01-0.97)
Ref
|
0.047
|
HIV Infection
Yes
No
|
2.53 (0.82-7.78)
Ref
|
0.11
|
2.55 (0.78-8.4)
Ref
|
0.123
|
HBV Infection
Yes
No
|
0.93 (0.24-3.58)
Ref
|
0.91
|
-
|
-
|
Hosmer Lemeshow Test for Multivariate Logistic regression, p=0.656 → Goodness of fit present
a Crude Odds Ratio (COR) b Adjusted Odds Ratio (AOR)
Association between MICS and nutritional, inflammatory, and other laboratory parameters
Patients with MICS had significantly lower mean dry-weight, BMI, haemoglobin, albumin, total cholesterol, transferrin, and creatinine levels. The median CRP level was significantly higher in patients with MICS. (Table 4)
Table 4. Association between MICS and nutritional, inflammatory and other laboratory parameters
Parameters
|
N
|
MICS Present (Mean + SD)
|
MICS Absent (Mean + SD)
|
p-value
|
Nutritional parameters
|
Dry weight (kg)
|
160
|
56.6 + 9.4
|
68.9 + 10.9
|
<0.01 a
|
BMI (kg/m2)
|
160
|
20.2 + 2.7
|
24.4 + 3.8
|
<0.01 a
|
Albumin (g/dl)
|
160
|
3.47 + 0.49
|
3.94 + 0.36
|
<0.01 a
|
Total Cholesterol (mg/dl)
|
160
|
138.7 + 34.4
|
163.3 + 40.6
|
<0.01 a
|
Inflammatory markers
|
Transferrin (mg/dl)
|
160
|
164.1 + 40.2
|
217.8 + 44.4
|
<0.01 a
|
CRP (mg/l)
|
103
|
22.5 (IQR=12.7-43.3)
|
10.45 (IQR=5.0-16.8)
|
<0.01b
|
Ferritin (ng/ml)
|
102
|
138.2 (IQR=75-349.8)
|
100.2 (IQR=49.5-216.7)
|
NS b
|
Other laboratory parameters
|
WBC (x 109/L)
|
160
|
5.21 + 2.78
|
4.93 + 1.79
|
NS a
|
Haemoglobin (g/dl)
|
160
|
8.8 + 2.0
|
9.7 + 1.8
|
<0.01 a
|
Pre-dialysis Cr (µmol/l)
|
160
|
768 + 316
|
911 + 412
|
0.016 a
|
Post-dialysis Cr (µmol/l)
|
160
|
267 + 128
|
361 + 202
|
<0.01 a
|
Urea Reduction Ratio (URR)%
|
160
|
69.9 + 13.6
|
70.4 + 9.0
|
NS a
|
a Test Statistic: Student’s t-test, b Test Statistic: Mann-Whitney U test
NS: Not significant (p>0.05), IQR-Interquartile Range
In this cross-sectional study, 160 haemodialysis patients were recruited to determine the prevalence of Malnutrition Inflammation Complex Syndrome and its associated factors at MNH in Dar es Salaam, Tanzania. More than two-thirds of the patients, 111 (69.4%) were male and the majority 132 (82.5%) were on HD three times per week (as per KDOQI recommendations). The prevalence of MICS was 46.3% and based on severity; MICS was categorized as mild in 24.4% and moderate to severe in 21.9% of the patients. In multivariate analysis, longer duration on HD (>4years) was a significant predictor of MICS and hypercholesterolaemia was a negative predictor of MICS. Patients with MICS had significantly lower mean BMI, albumin, total cholesterol, transferrin, haemoglobin, and creatinine levels.
The prevalence of MICS found in this study may not reliably reflect the situation in Tanzania due to rural-urban disparities in access to dialysis services [3]. A lower prevalence of MICS was noted in this study as compared to 61.2% reported by Matiko in a study conducted among HD patients in Kenya [11]. This difference could be attributed to the adequacy of dialysis as indicated by lower frequency (two times weekly) of dialysis in the study by Matiko as compared to our study in which most patients were getting three times per week dialysis. [11]. Inadequate dialysis can lead to persistently high urea levels, subsequent inflammation, and protein catabolism [12]. However, in this study, HD adequacy as assessed by URR did not associate with MICS. In this study, moderate to severe MICS was noted in 21.9% of HD patients, this is similar to a study from South Africa that reported it to be 22% [24].
The prevalence of underweight HD patients was 16.9%; this was higher than the prevalence of 9.8% reported by Somji et al in a study conducted in Dar es Salaam, Tanzania among HD patients [5]. The prevalence of underweight HD patients found in our study was lower when compared to other studies in African countries; 20.9% in Kenya [11], 29.2% in Niger [25], and 28.3% in Cameroon [26]. Patients with MICS have varying degrees of muscle wasting [12, 17]. In this study, patients with MICS had a mean BMI of 20.2 + 2.7 kg/m2 and almost half (48.4%) of the patients with normal BMI had MICS, this implies that BMI alone is not accurate for nutritional assessment in HD patients. The Malnutrition Inflammation Score is a better tool in the assessment of the nutritional status of HD patients and has been found to better correlate with morbidity, quality of life, and mortality [7, 9, 10]. Several factors affect the nutritional status of HD patients; decreased calorie intake, inadequate dialysis, metabolic acidosis, persistent inflammation, and HD-related catabolism [8, 12]. In this study, almost half of the patients had reported some deterioration in dietary intake and almost a third of patients reported gastrointestinal symptoms that may limit adequate dietary intake.
In our study, inflammation was present in 81.6% of the patients; this was lower than that found in the Kenyan study where all patients had raised CRP [11]. Inflammation in HD patients is common and largely attributed to persistent uremia. There are multiple other triggers including HD procedure itself, dialysate quality, and bio-incompatibility of dialysis membranes [4, 9]. The median CRP in our study was significantly higher in patients with MICS, this is consistent with a study done in Mexico [27].
In this study, patients on HD for a longer duration (>4 years) had 5 times the odds of having MICS compared to those on HD for less than 4 years. This finding is consistent with findings from a study by Omari et al who reported a long duration of dialysis to be linked with malnutrition inflammatory syndrome in Palestine [28]. The longer duration on HD therapy results in prolonged exposure of patients to negative effects of HD including dialysis-related inflammation, loss of nutrients, and increased energy expenditure that subsequently can result in malnutrition [12]. This highlights the need for efforts to increase uptake of HD patients into the kidney transplantation program to limit the time spent on chronic HD therapy.
Low socioeconomic status and limited access to health insurance significantly influence the adequacy of haemodialysis [4]. We found that the small group of patients who were paying ‘out of pocket’ (ie. not insured) had slightly higher odds of having MICS. There was a significant difference in the frequency of HD among these two groups; 87.5% of those paying ‘out of pocket’ versus 9.7% of those who had health insurance were having twice per week HD (p<0.05). Thus health insurance offers the best option for sustaining the adherence to standard long-term haemodialysis.
Aging has been reported to correlate with MICS but did not show this effect in our study. This might be attributed to the younger age of patients in this study compared to that reported by Rambod et al in a study among HD patients in California who found a higher burden of MICS among older patients on HD [9]. Similarly, other studies on MICS done in Kenya, Brazil, and Mexico also had a younger population; mean age <50years [10, 11, 27]. Some of the possible reasons could be due to variations in the profile of the HD patients which can be affected by differences in survival, recruitment in renal transplantation, and the access to sustainable maintenance HD services. MICS among older patients is attributed largely to co-morbidities including depression that may also affect the dietary intake of these patients.
In this study, the mean total cholesterol level was significantly lower in patients with MICS, which is contrary to a report by Valencia et al that reported no difference in cholesterol levels between dialysis patients with and without MICS in Mexico [27]. Cholesterol is an important nutritional biomarker among patients on HD therapy and hypercholesterolaemia may be protective in these patients unlike the general population [29, 30].
Serum albumin was noted to be significantly lower among patients with MICS; this was an expected finding consistent with reports from other similar studies [9, 10]. Albumin is a good indicator of protein-energy malnutrition in CKD. Hypoalbuminaemia is attributed to visceral protein loss among patients undergoing HD as well as peritoneal dialysis [12]. Serum creatinine levels have also been reported to be lower among patients with MICS. [10, 27].
The phenomenon of hypocholesterolaemia and hypocreatininaemia among patients on chronic dialysis, accompanied by MICS is referred to as ‘reverse epidemiology’. Reverse epidemiology is paradoxically associated with poor cardiovascular outcomes [29, 30], therefore health care providers providing chronic dialysis therapy should be vigilant and monitor these markers regularly to improve outcomes of chronic dialysis.
Strengths and Limitations
This is the first study in Tanzania to explore the burden of malnutrition and inflammation in HD patients. The other strength of this study was that it also assessed the relationship between nutritional and inflammatory biomarkers with MICS and was shown to be correlated hence reinforces the need for synergistic assessment of these parameters.
In view of the study settings, the data may not be generalized for the whole country. Another limitation of this study was the observer bias while measuring fat and muscle loss. Systematic errors during the measurement of anthropometric indices could have occurred. There might be some selection bias from (i) larger number of health insured patients compared to the small number of patients paying ‘out of pocket’ (ii) exclusion of a small group of very sick and mentally debilitated patients. This may underestimate the outcome. The CRP test was not done for all patients; this was due to financial restrictions and other logistic challenges.