CHAPTER PART 2 OF 9
One table to write the diet from, from CKD G1 through haemodialysis and peritoneal dialysis, with every figure carrying the grade the guideline actually gives it.
One table to write the diet from. The grade chips are the guidelines' own, shown beside the number they belong to, so staff can tell a strong recommendation from an expert opinion at a glance.
kcal/kg/day, adjusted for age, sex, activity, body composition and inflammation. 1C
per day (90 mmol, under 5 g salt). KDOQI allows up to 2.3 g. Indian intake averages 11 g salt. 1B 2C
g/kg/day per KDOQI. KDIGO and ADA say a flat 0.8. See the disagreement below.
g/kg/day, haemodialysis and peritoneal dialysis alike. Higher if glycaemic swings demand it.
| Population | Protein g/kg/d | Energy kcal/kg/d | Sodium | Potassium | Phosphate | Fluid |
|---|---|---|---|---|---|---|
| CKD G1–G2 | 0.8 | 25–35 | <2 g 1B 2C | No restriction. Normal serum potassium means normal diet. | No restriction. Avoid additive phosphate. | 1.5–2 L if not oedematous. Do not coach extra water. |
| CKD G3–G5, not on dialysis | 0.6–0.8 OPINION or 0.8 2C | 25–35 1C | <2 g 1B 2C | Titrate to serum K. Target bioavailable and additive sources first. OPINION | Adjust to keep serum phosphate in range 1B. Weight the source, not the milligrams OPINION. | Individualise to urine output, oedema, blood pressure. No guideline number exists. |
| Very low protein, supervised NON-DIABETIC CKD ONLY |
0.28–0.43 + ketoanalogues | 30–35 | <2 g | As above | Falls with protein | As above |
| Maintenance haemodialysis | 1.0–1.2 OPINION | 25–35 | <2 g, often 1–1.5 g if hypertensive | Titrate to pre-dialysis serum K. Most need real restriction. | Binder-led. Match binder dose to the phosphate in each meal. | Governed by interdialytic weight gain: target under 4 to 4.5 percent of body weight. |
| Peritoneal dialysis | 1.0–1.2 OPINION, higher in peritonitis | 25–35 minus dialysate glucose | <2 g | Often normal or low. Do not restrict reflexively. | Binder-led | Set by ultrafiltration and residual urine, not a fixed allowance. |
| Acute illness or hospitalised, any stage | 1.0 rising to 1.3; on CRRT 1.5–1.7 | 30–35 | Individualise | Biochemistry-led, daily | Biochemistry-led | Clinical |
The 0.28 to 0.43 g/kg figure comes from KDOQI Statement 3.0.1, which is explicitly the non-diabetic CKD statement, and KDIGO's very low protein practice point is not diabetes-specific either. The Cochrane review in diabetic kidney disease tested only 0.6 to 0.8 against 1.0 or more. Nothing supports a very low protein diet in diabetic CKD.
If a nephrologist prescribes one anyway, it requires ketoanalogue supplementation, monthly review with weight, handgrip and albumin, and a specific reassessment of the SGLT2 inhibitor, because the combination of a very low protein intake, a low carbohydrate intake and an SGLT2 inhibitor is a euglycaemic ketoacidosis setup.
For CKD G3 to G5 with diabetes, KDOQI 2020 permits 0.6 to 0.8 g/kg/day, graded OPINION. KDIGO 2022 and 2024, and ADA, hold at 0.8 g/kg/day, graded 2C and A respectively (the verbatim grade A text is ADA 2025 Recommendation 11.8; the 2026 edition is not thought to have changed the numbers). There is no trial evidence discriminating between them: the Cochrane review of low protein diets in diabetic kidney disease found only 8 studies and 486 participants, with uncertain effects on death and end-stage disease.
House position for this clinic: write 0.8 g/kg/day as the default. Go to 0.6 only in a patient who is metabolically stable, not sarcopenic, not on a GLP-1 agonist, and who will be seen monthly with weight, handgrip and albumin tracked. Most Indian patients already eat 0.6 to 0.8 g/kg without being told to, so the more common error here is under-eating protein, not over-eating it.
About 60 percent of the glucose instilled in the dialysate is absorbed, which is roughly 100 to 200 g per day. In one series that was 20 percent of the patient's total energy intake, arriving continuously, unrecorded in any food diary, and driving both hyperglycaemia and weight gain. Subtract it before you write the oral energy prescription, and expect the diabetes to look worse on PD than it did before.