A prospective cohort study of prevalent and incident PD patients in a dialysis unit of Federal University Hospital in Brazil between July 2021 and July 2022.

The inclusion criteria were age ≥18 years and a signed informed consent form (ICF)

The exclusion criteria were not signed ICF or age below 18 years (Figure 1).

Data collection procedure

We evaluated the sociodemographic and clinical data. Adherence to the technique was assessed by means of attendance at medical appointments, ultrafiltration (UF) records and body weight in the monthly dialysis reports, as well as correct information on the number of dialysis bags checked by nurses. To be considered adherent, the patient should be within the proposed goals for at least 80% of the follow uptime.

Laboratory tests were also recorded according to the National Health Surveillance Agency (ANVISA) protocol¹⁰. The reference values considered were fasting glucose (< 110 mg/dl), hemoglobin (Hb) (> 10 g/dl), transferrin saturation index (TSI) (> 20%), ferritin (100-600 ng/dl), calcium (8.5-10.5 mg/dl), phosphorus (< 5.5 mg/dl), parathyroid hormone (PTH) (150-600 pg/ml), alkaline phosphatase (ALP) (65-300 U/L), potassium (< 5.5 mEq/L), albumin (> 3.5 g/dl) and weekly Kt/V (> 1.7). Thus, we evaluated a total of 13 adequacy goals. An annual peritoneal equilibrium test (PET)¹¹ and an evaluation of urea kinetic adequacy f (Kt/V)¹² were performed. Adequacy to the clinical and laboratory goals follows the same definitions to the technique adherence that is the patient should be within the proposed goals for at least 80% of the follow uptime.

The patients also underwent the application of scores for the analysis of health literacy by the Short Assessment of Health Literacy for Portuguese-speaking adults score (SAHLPA-18)¹³, a Scale of Perception of Social Support (SPSS)¹⁴ , the assessment of nutritional status by a Malnutrition Inflammation Score (MIS)¹⁵ and a tool for the evaluation of QoL (SF-12)¹⁶.

The main drugs used were recorded. The following complications of the method were evaluated: exit site infection (ESI) and tunnel infections, peritoneal dialysis associated peritonitis (PD peritonitis) and mechanical complications (hernias, leakages, fistulas and catheter translocations) according to the criteria of the International Society of Peritoneal Dialysis (ISPD)¹⁷.

Death and QoL were defined as primary, whereas PD discontinuation and hospitalization were defined as secondary.

We performed a descriptive analysis and subsequently, we compared incidents (patients who started PD after the first day of July 2021) versus prevalent (patients who were in PD before the first day of July 2021) with respect to sociodemographic, clinical, laboratory, scores, complications and outcomes. This same analysis was performed between patients in and out goals in order to verify the association of these variables with goals.

We consider as goals the achievement of a total of thirteen clinical and laboratory parameters of dialysis adequacy (BP, edema, Hb, TSI, ferritin, blood glucose, calcium, phosphorus, ALP, PTH, potassium, albumin and Kt/V).

The secondary and primary outcomes were correlated with the 13 established adequacy goals to verify which factors interfered with them. Finally, we used a linear regression model to evaluate the variables associated with QoL (physical and mental components) and logistic regression to evaluate outcomes: hospitalization, PD discontinuation and death; these models were adjusted for confounders with biological plausibility and/or statistical significance. We used Statistical Package for the Social Sciences 25.0 Chicago, Illinois, and adopted a CI of 95%.

The study was approved by the Ethics and Research Board of the University Hospital of the Federal University of Juiz de Fora under no 5.001.893. All patients signed ICF, STROBE rules and the Declaration of Helsinki were followed.

All the 56 patients evaluated were in automated peritoneal dialysis (APD) because no patients chose the continuous ambulatory peritoneal dialysis (CAPD) modality. Furthermore, all of them use glucose-based solutions (icodextrin is not available in the public health system in Brazil).

In total, 73.2% (41 patients) being prevalent and 26.8% (15 patients) incident patients. These patients were on Home Choice APD, but not Claria system™(Baxter) with the Sharesource™ (Baxter) remote patient management (RPM). The prescriptions were made by the attending nephrologist and adjusted according to the clinical and laboratory conditions of the patient. Twenty-seven percent of the patients were on incremental dialysis during the study (which means less than seven days a week performing PD). The majority were women, with a mean age of 61.2 ± 13.9 years, white, illiterate and with low income (Table I). Only two patients did not have caregivers (Table Ia andTable Ib).

Among the underlying CKD, the main cause was undetermined. All patients were hypertensive and one-third were diabetic A total of 76.8% of the patients were considered adherent to the method. Most patients had previous nephrological follow-up (). Of the prevalent patients, 22% had already presented at least one episode of peritonitis, and regarding previous KRT, 67.9% had undergone HD for 23 ±4 months on average.

The SAHLPA¹³ revealed low health literacy. Regarding the MIS¹⁶, the values were found to be low (Figure 2.1). According to the SF-12¹⁵ scores, the total average scores for the physical component and the mental component were low (Figure 21). The SPSS¹⁴ showed an adequate total mean in its practical and emotional components (Figure 2.3). According to the BMI⁹, 50.9% were at their ideal weight (Table Ia-Ib).

Among the main drugs used, angiotensin converter enzyme inhibitors (ACEIs) or angiotensin receptor blockers (ARBs) and erythropoietin (Epo) were the most used (Figure 2.4).

Among the 73.2% patients who underwent PET¹¹, 9.7% had slow peritoneal solute transfer rate (PSTR), 53.6% had medium-low PSTR, 34.1% had medium-fast PSTR and only 4.1% had fast PSTR. When Kt/V¹² was performed, the overall mean was 1.89 ± 0.61.

Among the patients who underwent echocardiography (75% of the sample), 81% had some degree of left ventricular hypertrophy, which was considered moderate or severe in 40.5%. On radiological analysis, a pattern of pulmonary venocapillary congestion was observed in one-third of the sample, with a statistically significant difference between incidents (7.7%) and prevalent cases (41.4%) (p = 0.04).

The median diuresis of the study population was 814 ml and was higher among incident patients (Table Ia¹³). Nine patients in the study were anuric, eight of whom were prevalent cases and one of whom was an incident case. This incident patient underwent HD previously.

Regarding the adequacy goals in PD we noted that, in general, the goals of blood pressure (BP) (30.4%) and phosphorus (28.6%) were the most difficult to achieve. On the other hand, the goals most commonly achieved were potassium (62.5%), albumin (60.7%), KtV¹² (61.5%) and the absence of edema (60.7%). It’s important to note that 100% of the patients had a diagnosis of SAH but 69.6% remained out of the goals during the study even without edema. When evaluating the hematological parameters, Hb (42.9%), TSIs (39.3%) and ferritin (35.7%), we observed a minority of patients within the goals. When bone mineral disorder (BMD-CKD) was assessed, calcium was found to be a low therapeutic target (37.5%), while PTH (51%) and ALP (50%) were the best targets. In the comparison between prevalent and incident patients, none of these variables showed statistical significance (Figure 3).

Adequacy goals in PD

BP- blood pressure; Edema- presence of “Locker” in the lower limbs; Hb-hemoglobin; TSI- transferrin saturation index; Ca-calcium; P-phosphorus; ALP-alkaline phosphatase; PTH- parathyroid hormone; K-potassium; KtV- kinetic urea removal index; MIS- malnutrition and inflammation score; SF-12- Short Form-12 quality of life score; SPSS- Scale of Perceived Social Support; ACEIs/ARBs- angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers.

Analysis of the factors that interfered with the goals, as shown in , we found that for the BP goal, advanced age (p=0.03), lower adherence to the technique (p = 0.005) dementia (p=0.04), the presence of peritonitis (p=0.05) and a worst mental component of the SF-12¹⁵ (p=0.048) were associated with the worst goals (Figure 4).

Among the hematological parameters, Hb was more on the target among those with previous nephrological follow-up (p = 0.01), lower MIS¹⁶ (p = 0.002) and a greatest physical component of the SF-12¹⁵ (p = 0.006). The best ferritin levels and the use of ACEIs/ARBs (p = 0.03) were associated with a lower prevalence of dementia (p = 0.04). In turn, TSIs were related only to hospitalizations, and conversely, a higher level of TSIs was associated with a higher rate of hospitalizations (p = 0.05). This association is not easily explained and is possible because of the small sample (Figure 4).

In patients for whom phosphorus control was more common at the goal, we observed older age (p=0.04), etiologies other than autosomic dominant polycystic kidney disease (p = 0.02) and the presence of congestive heart failure (p = 0. 03) and, paradoxically, the lack of previous nephrological follow-up (p = 0.05). Calcium levels were better controlled in brown patients (p = 0.04), those with lower MIS¹⁶ (p = 0.05), those with lower practical SPSS¹⁴ (p = 0.04), and those with lower hospitalization rates (p = 0.04). PTHi was better controlled in nonMIS-malnutritionandinflammationscore;SAHLPA-bese individuals (p = 0.02) and among those who had not undergone previous HD (p = 0.04) (Figure 4). In the evaluation of glucose, diabetic patients were more likely to be out of target (p = 0.001), and those who were younger (p = 0.005), were without dementia (p = 0.05), had more previous nephrological follow-up (p = 0.05) or had a worse physical component of the SF-12¹⁵ (p = 0.017) were among those who were on target (Figure 4).

Patients who met the potassium goal had a higher income (p = 0.03), nondiabetic CKD etiology (p = 0.003), and less progress to death (p = 0.05) .

Albumin achieved better results in non-dementia patients (p = 0.04), in those who did not undergo previous PD (p = 0.04), and in those who did not have peritonitis (p = 0.02). Patients with higher KtV¹² levels were more adherent (p = 0.008) and had a lower percentage of hospitalizations (p = 0.04) (Figure 4).

Example caption for this image.

We observed relative heterogeneity in the determinants of the goals that may be due to the sample size. However, several factors, such as age, dementia status, DM status, and previous nephrological follow-up, strongly interfered with several goals (Figure 4).

Regarding the secondary outcomes, 32 patients were admitted to the hospital (57.1%), 20 of whom were prevalent (48.8%) and 12 of whom were incident (80%), revealing that there were more hospitalizations among incident cases (p = 0.03). Among those of the 39 hospitalizations, the main etiology responsible for hospitalization was sepsis unrelated to PD (12 patients or 30.8%), notably due to respiratory, urinary tract and skin infections.

ESI occurred in 18 cases (32.1%), 10 of which were prevalent cases (24.4%) and eight of which were incident patients (53.3%), and there were no tunnel infections. There was a higher proportion of ESI in incident patients (p = 0.05). Peritonitis occurred in seven patients (12.5%), four of whom were prevalent patients (9.8%) and three of whom were incident patients (20%) (p = 0.37).

Eight individuals (14,3%) were affected by mechanical complications, including leakage (4), hernias (2), catheter translocation (1), and pleuroperitoneal fistula (1). These events occurred in seven incidents (46.6%) but in only one prevalent patient (2.4%) (p = 0.001)

.

Changes in technique occurred in 10 patients (17.8%), one who received a transplant (1.8%) and nine who were referred for HD (16%). One third of the incident patients were transferred to HD, which was a higher percentage than that of prevalent patients (9.8%) (p = 0.09). The main reasons were poor adherence to therapy (4), peritonitis (3), pleuroperitoneal fistula (1) and UF failure (1).

During follow-up, there were 23.2% deaths, 26.8% of which were prevalent and 13.3% of which were incident patients. It is noteworthy that the deaths among incident cases occurred in those with less than three months of therapy. Sepsis unrelated to PD was identified as the main cause of death in eight cases (62%), six of which were prevalent and two were incident patients. There were three deaths due to cardiovascular disease (23%) and two deaths due to coronavirus disease (COVID-19) (15.4%), all of which were prevalent patients. It should be noted that there was no record of death due to sepsis associated with PD.

Nevertheless, regarding secondary outcomes within the multivariate analysis model with the outcome variable of hospitalization, higher diuresis were associated with more hospitalization and those patients without diabetes presented lower incidence of hospitalization Table II. Withregard to the multivariate analysis, it was not possible to find a model with adequate analysis of residues that correlated PD discontinuation with other variables.

We evaluated the primary outcomes in a multivariate model after performing multiple adjustments for confounders, with the outcome variable QoL analyzed separately for its physical and mental components and death (Table III)

In a linear regression model, with the physical component of the SF-12 (PCS-12) ¹⁵as the dependent variable, we observed that only the MIS ¹⁶ score showed a negative correlation. In the analysis of the mental component of the SF-12 (MCS-12) as the dependent variable, only age was negatively correlated . There was no significant difference in the outcome of death.