Study design, duration and setting
It is a multi-center (two HD centers of MNH) prospective cross-sectional observational study. The study was conducted between September and November 2019 among patients undergoing haemodialysis at the two HD centers of Muhimbili National Hospital (Upanga and Mloganzila) in Dar es Salaam, Tanzania.
Study population
All adult patients receiving maintenance HD therapy at the two dialysis units of MNH for at least three months were eligible. Patients who were mentally incapacitated, very sick, bedridden patients, patients with altered level of consciousness and those in respiratory distress were excluded.
Sample size calculation
The sample size for this study was estimated based on the prevalence of MICS of 61.2% reported by Matiko among HD patients in Kenya, using the following formula below [18].
n = no x N where, no = Z2 p (1-p)
no + (N-1) e2
Whereby:
n = Minimum sample size needed
no = Is the sample size without considering the finite population correction factor
N= The total number of haemodialysis patients at both renal centers of MNH (At the time of preparation of the study protocol there were 284 patients on HD)
Z = Standard normal deviation = 1.96 (at 95% Confidence Interval)
p = Prevalence of MICS. p=61.2%=0.612 among HD patients in a study done in Kenya [11].
e = Margin of error which was set at 5%=0.05
Substituting these figures in the formula gives; n = 160
Therefore, the minimum required sample size was 160 patients.
Sampling method
Simple random sampling was utilized to select participants for this study. A list of all patients on HD therapy was obtained from the HD units record books at the beginning of the study. Patients were assigned numbers then the Stat Trek's Random Number Generator was used to select participants recruited in this study [19].
Data collection methods
Data were actively collected by an ad hoc questionnaire which included participant’s demographic and clinical data, questions about dietary intake, gastrointestinal symptoms; functional capacity, and co-morbid status. These were taken from the Malnutrition Inflammation Score (MIS) [7]. The Malnutrition Inflammation Score is a comprehensive assessment of the nutritional and inflammation status of patients. It has 10 components (5 from medical history, 3 from physical examination and 2 from laboratory parameters: albumin and transferrin) each has a severity ranging from 0 (normal) to 3 (severely abnormal). The total of all 10 components ranges from 0 (normal) to 30 (severely malnourished); higher scores imply more severe degree of malnutrition and inflammation. The MIS components include (i) Change in end dialysis dry weight over past 3 months (ii) Dietary intake (iii) Gastrointestinal symptoms (iv) Functional capacity (v) Co-morbidity including the number of years on dialysis (vi) Loss of fat stores (vii) Signs of muscle wasting (viii) Body mass index (BMI) (ix) Serum albumin and (x) Serum Total Iron binding capacity (TIBC) or serum transferrin. Dietary intake was scored 0 for the usual intake of solid foods, with no recent decrease in the amount/quality of meals. A slightly suboptimal solid diet was scored 1, a full-liquid diet or moderate decrease in dietary intake was scored 2 and a score of 3 indicated a hypocaloric liquid diet or starvation. The full description and grading of the severity of these components is available in the study by Kalantar-Zadeh et al, who is the inventor of this scoring system [7].
Body (or somatic) mass was assessed using the body mass index (BMI) that was determined using participants’ post-dialysis dry weight and height. BMI was computed using the weight (kg) and height (m) and was expressed in kg/m2: BMI = Weight (kg) ÷ [Height (m)] 2
Body weight was measured using a standard weighing scale while the patient had not worn shoes and a stadiometer was used to measure the height of participants to the nearest centimeter. For patients who were unable to stand in an upright position on the stadiometer, the length from top of the head to the plantar surface of the foot while lying supine was taken in lieu of their height.
Patients’ dry weight in the preceding three months was obtained from their HD records and were compared with current post-dialysis dry weight to calculate the percentage change. Loss of subcutaneous fat was assessed below the eyes, triceps, biceps, and chest. Signs of muscle wasting were assessed at the temple, clavicle, scapula, ribs, quadriceps, knee, and interosseous regions. Subcutaneous fat loss, muscle wasting and BMI were categorized based on the MIS [7].
Laboratory tests
Blood specimen (10 mls) was obtained from each participant before and after dialysis, and was sent to the Central Pathology Laboratory at MNH. Laboratory testing was performed to determine complete blood count (CBC), C-reactive protein (CRP), ferritin, transferrin, creatinine, urea, total cholesterol and albumin. Post-dialysis creatinine and urea were also determined. CELL DYN 3700 and ARCHITECT PLUS machines were used to analyze complete blood count and serum biochemistry respectively. CRP was analyzed at Muhimbili University of Health and Allied Sciences (MUHAS) Clinical Research Laboratory using the COBAS INTEGRA 400, which uses Finecare CRP Rapid Quantitative test that is based on the fluorescence immunoassay.
Data management and analysis
Questionnaires were checked for completeness after which data were entered into the Statistical Package for Social Sciences (SPSS) version 20 for data cleaning and analysis. Data were summarized into frequency distribution and two-way tables. Association between categorical variables were determined using Chi square and Fischer’s exact tests while continuous variables association was determined using student t test and analysis of variance (ANOVA). Non-parametric variables were compared using Mann-Whitney U test. Binary Logistic regression analysis (univariate and multivariate) was performed to determine the predictors/factors associated with MICS. A two-tailed p-value of less than 0.05 was regarded as statistically significant and the Hosmer-Lemeshow test was used to determine the goodness of fit for the final logistic regression model.
Study variables
The main outcome variable was presence of MICS as defined by an aggregate MIS of 6 or more. The severity of MICS was graded as: A MIS of less than 6 was categorized as Normal, between 6 and 10 as Mild and a score of 11 or more was considered as Moderate to Severe [11]. Body size (BMI) categories for this study were defined as Underweight (BMI < 18.5 kg/m2), Well-nourished (BMI 18.5 to 24.99 kg/m2) and Overweight or Obese (BMI > 25 kg/m2) [20]. A pre-dialysis CRP level above 5 mg/l was defined as having Inflammation [21]. Anemia was defined by a haemoglobin level below 13 g/dl in males and below 12 g/dl for females based on KDIGO guidelines [22]. Hypercholesterolaemia was defined by a serum total cholesterol (TC) level above 220 mg/dl (the upper limit of the normal range at MNH Laboratory). Adequate dialysis was defined as having Urea Reduction Ratio (URR) of 65% or more (calculated as URR = (Pre-dialysis Urea - Post-dialysis Urea)/Pre-dialysis Urea X 100%) [23].