diff --git a/README.md b/README.md
index f7016af5..bae3ae03 100644
--- a/README.md
+++ b/README.md
@@ -187,7 +187,7 @@ echo "SELECT * FROM v_master LIMIT 5;" | docker exec -i supabase_db_tryvit psql
 ```
 ┌─────────────────┐     ┌──────────────────┐     ┌─────────────────────────┐
 │  Open Food Facts │────▶│  Python Pipeline │────▶│  PostgreSQL (Supabase)  │
-│  API v2          │     │  sql_generator   │     │  231 migrations         │
+│  API v2          │     │  sql_generator   │     │  232 migrations         │
 │  (category tags, │     │  validator       │     │  58 pipeline folders    │
 │   countries=PL,DE│     │  off_client      │     │  products + nutrition   │
 └─────────────────┘     └──────────────────┘     │  + ingredients + scores │
@@ -319,7 +319,7 @@ tryvit/
 │   └── views/                       # Reference view definitions
 │
 ├── supabase/
-│   ├── migrations/                  # 231 append-only schema migrations
+│   ├── migrations/                  # 232 append-only schema migrations
 │   ├── seed/                        # Reference data seeds
 │   ├── tests/                       # pgTAP integration tests
 │   └── functions/                   # Edge Functions (API gateway, push notifications, CAPTCHA)
@@ -531,3 +531,4 @@ This project is licensed under the terms in the [LICENSE](LICENSE) file.
   <br />
   <em>Built with science and care.</em>
 </p>
+
diff --git a/copilot-instructions.md b/copilot-instructions.md
index e68f689a..0a87ee04 100644
--- a/copilot-instructions.md
+++ b/copilot-instructions.md
@@ -189,7 +189,7 @@ tryvit/
 │   │   ├── api-gateway/             # Write-path gateway (rate limiting, validation) (#478)
 │   │   └── send-push-notification/  # Push notification handler
 │   ├── dr-drill/                    # Disaster recovery drill artifacts
-│   └── migrations/                  # 231 append-only schema migrations
+│   └── migrations/                  # 232 append-only schema migrations
 │       ├── 20260207000100_create_schema.sql
 │       ├── 20260207000200_baseline.sql
 │       ├── 20260207000300_add_chip_metadata.sql
@@ -698,7 +698,7 @@ a mix of `'baked'`, `'fried'`, and `'none'`.
 
 ## 7. Migrations
 
-**Location:** `supabase/migrations/` — managed by Supabase CLI. Currently **231 migrations**.
+**Location:** `supabase/migrations/` — managed by Supabase CLI. Currently **232 migrations**.
 
 **Rules:**
 
@@ -2705,3 +2705,4 @@ gh issue comment XXX --body "Session complete — <what was done, what's next>"
 - Sections are numbered sequentially so any agent can say "per §X.Y" unambiguously.
 - When this file is updated, the commit message should include the sections changed (e.g., `chore(docs): copilot-instructions §4 — add ingredient_translations table`).
 - When the instructions conflict with a specific issue's acceptance criteria, the issue wins for that specific change — but document the divergence in the PR description.
+
diff --git a/db/ci_post_enrichment.sql b/db/ci_post_enrichment.sql
index 12252256..cd78d3e7 100644
--- a/db/ci_post_enrichment.sql
+++ b/db/ci_post_enrichment.sql
@@ -197,32 +197,150 @@ AND NOT EXISTS (
 )
 AND EXISTS (SELECT 1 FROM product_ingredient LIMIT 1);
 
+-- ═══════════════════════════════════════════════════════════════
+-- Step 0e: Classify additive/E-code concern tiers
+-- ═══════════════════════════════════════════════════════════════
+-- Uses strict start-of-string E-code extraction to avoid false positives
+-- such as "200 G", "Type 500", or nutrition text.
+
+WITH normalized AS (
+  SELECT
+    ingredient_id,
+    name_en,
+    CASE
+      WHEN lower(name_en) ~ '^\s*e[ -]?[0-9]{3,4}[a-z]?($|[^a-z0-9])'
+      THEN regexp_replace(
+        substring(lower(name_en) from '^\s*e[ -]?[0-9]{3,4}[a-z]?'),
+        '[^a-z0-9]',
+        '',
+        'g'
+      )
+      ELSE ''
+    END AS e_code,
+    lower(name_en) AS n
+  FROM ingredient_ref
+  WHERE name_en IS NOT NULL
+),
+classified AS (
+  SELECT
+    ingredient_id,
+    CASE
+      WHEN e_code IN ('e250','e252')
+        OR n LIKE '%sodium nitrite%'
+        OR n LIKE '%potassium nitrate%'
+      THEN 3
+
+      WHEN e_code IN ('e133','e150d','e211','e220','e223','e319','e385','e407','e407a','e621','e950','e951','e955')
+        OR n LIKE '%sodium benzoate%'
+        OR n LIKE '%benzoate%'
+        OR n LIKE '%sucralose%'
+        OR n LIKE '%aspartame%'
+        OR n LIKE '%carrageenan%'
+        OR n LIKE '%sodium metabisulphite%'
+        OR n LIKE '%contains sulphites%'
+        OR n IN ('sulfite','sulfiten','sulphite','sulphites')
+      THEN 2
+
+      WHEN e_code IN (
+        'e150','e150a','e150b','e150c','e172','e200','e202',
+        'e281','e282','e338','e339','e340','e341',
+        'e420','e421','e422','e440','e450','e451','e452',
+        'e460','e466','e471','e472','e500','e501','e503',
+        'e960','e960a','e960c'
+      )
+        OR n LIKE '%potassium sorbate%'
+        OR n LIKE '%sorbate%'
+      THEN 1
+
+      ELSE 0
+    END AS target_concern_tier,
+    CASE
+      WHEN e_code IN ('e250','e252')
+        OR n LIKE '%sodium nitrite%'
+        OR n LIKE '%potassium nitrate%'
+      THEN 'Strict additive/E-code classifier: nitrite/nitrate preservative concern'
+
+      WHEN e_code IN ('e133','e150d','e211','e220','e223','e319','e385','e407','e407a','e621','e950','e951','e955')
+        OR n LIKE '%sodium benzoate%'
+        OR n LIKE '%benzoate%'
+        OR n LIKE '%sucralose%'
+        OR n LIKE '%aspartame%'
+        OR n LIKE '%carrageenan%'
+        OR n LIKE '%sodium metabisulphite%'
+        OR n LIKE '%contains sulphites%'
+        OR n IN ('sulfite','sulfiten','sulphite','sulphites')
+      THEN 'Strict additive/E-code classifier: preservative, sweetener, sulphite, or carrageenan concern'
+
+      WHEN e_code IN (
+        'e150','e150a','e150b','e150c','e172','e200','e202',
+        'e281','e282','e338','e339','e340','e341',
+        'e420','e421','e422','e440','e450','e451','e452',
+        'e460','e466','e471','e472','e500','e501','e503',
+        'e960','e960a','e960c'
+      )
+        OR n LIKE '%potassium sorbate%'
+        OR n LIKE '%sorbate%'
+      THEN 'Strict additive/E-code classifier: lower concern additive or sorbate'
+
+      ELSE NULL
+    END AS target_concern_reason
+  FROM normalized
+)
+UPDATE ingredient_ref ir
+SET
+  concern_tier = GREATEST(COALESCE(ir.concern_tier, 0), c.target_concern_tier),
+  is_additive = true,
+  concern_reason = CASE
+    WHEN ir.concern_reason IS NULL OR trim(ir.concern_reason) = ''
+    THEN c.target_concern_reason
+    ELSE ir.concern_reason
+  END
+FROM classified c
+WHERE ir.ingredient_id = c.ingredient_id
+  AND c.target_concern_tier > 0
+  AND (
+    COALESCE(ir.concern_tier, 0) < c.target_concern_tier
+    OR ir.is_additive IS DISTINCT FROM true
+    OR ir.concern_reason IS NULL
+    OR trim(ir.concern_reason) = ''
+  );
+
 -- ═══════════════════════════════════════════════════════════════
 -- Step 1: Populate ingredient_concern_score from actual ingredient data
 -- ═══════════════════════════════════════════════════════════════
 -- Based on EFSA concern tiers: tier 1 = 15pts, tier 2 = 40pts, tier 3 = 100pts
 -- Capped at LEAST(100, SUM(...)) per SCORING_METHODOLOGY.md v3.2
 
+WITH product_scores AS (
+  SELECT
+    pi.product_id,
+    LEAST(100, SUM(
+      CASE ir.concern_tier
+        WHEN 1 THEN 15
+        WHEN 2 THEN 40
+        WHEN 3 THEN 100
+        ELSE 0
+      END
+    ))::int AS score
+  FROM product_ingredient pi
+  JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
+  WHERE COALESCE(ir.concern_tier, 0) > 0
+  GROUP BY pi.product_id
+),
+computed_scores AS (
+  SELECT
+    p.product_id,
+    COALESCE(ps.score, 0)::int AS score
+  FROM products p
+  LEFT JOIN product_scores ps ON ps.product_id = p.product_id
+  WHERE p.is_deprecated IS NOT TRUE
+)
 UPDATE products p
-SET ingredient_concern_score = COALESCE(concern.score, 0)
-FROM (
-    SELECT pi.product_id,
-           LEAST(100, SUM(
-               CASE ir.concern_tier
-                   WHEN 1 THEN 15
-                   WHEN 2 THEN 40
-                   WHEN 3 THEN 100
-                   ELSE 0
-               END
-           ))::int AS score
-    FROM product_ingredient pi
-    JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
-    WHERE ir.concern_tier > 0
-    GROUP BY pi.product_id
-) concern
-WHERE concern.product_id = p.product_id
+SET ingredient_concern_score = cs.score
+FROM computed_scores cs
+WHERE p.product_id = cs.product_id
   AND p.is_deprecated IS NOT TRUE
-  AND p.ingredient_concern_score IS DISTINCT FROM COALESCE(concern.score, 0);
+  AND p.ingredient_concern_score IS DISTINCT FROM cs.score;
 
 -- ═══════════════════════════════════════════════════════════════
 -- Step 2: Flag palm oil controversy from actual ingredient data
diff --git a/db/qa/QA__ingredient_quality.sql b/db/qa/QA__ingredient_quality.sql
index ba4a5b35..aab243a2 100644
--- a/db/qa/QA__ingredient_quality.sql
+++ b/db/qa/QA__ingredient_quality.sql
@@ -198,3 +198,43 @@ SELECT '17. ingredient_translations source valid' AS check_name,
 FROM ingredient_translations
 WHERE source NOT IN ('curated', 'off_api', 'auto_translated', 'user_submitted');
 
+
+-- ═══════════════════════════════════════════════════════════════════════════
+-- 18. concern_tier should not remain all-zero once product_ingredient exists
+-- ═══════════════════════════════════════════════════════════════════════════
+SELECT '18. concern_tier populated when product ingredients exist' AS check_name,
+       CASE
+         WHEN NOT EXISTS (SELECT 1 FROM product_ingredient) THEN 0
+         WHEN EXISTS (
+           SELECT 1
+           FROM product_ingredient pi
+           JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
+           WHERE COALESCE(ir.concern_tier, 0) > 0
+         ) THEN 0
+         ELSE 1
+       END AS violations;
+
+-- ═══════════════════════════════════════════════════════════════════════════
+-- 19. products linked to concern-tier ingredients should have concern score
+-- ═══════════════════════════════════════════════════════════════════════════
+SELECT '19. concern-tier linked products have concern score' AS check_name,
+       CASE
+         WHEN NOT EXISTS (
+           SELECT 1
+           FROM product_ingredient pi
+           JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
+           WHERE COALESCE(ir.concern_tier, 0) > 0
+         ) THEN 0
+         ELSE (
+           SELECT COUNT(*)
+           FROM (
+             SELECT DISTINCT p.product_id
+             FROM products p
+             JOIN product_ingredient pi ON pi.product_id = p.product_id
+             JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
+             WHERE p.is_deprecated IS NOT TRUE
+               AND COALESCE(ir.concern_tier, 0) > 0
+               AND COALESCE(p.ingredient_concern_score, 0) = 0
+           ) x
+         )
+       END AS violations;
diff --git a/db/qa/QA__scoring_formula_tests.sql b/db/qa/QA__scoring_formula_tests.sql
index a815341a..22f5f37e 100644
--- a/db/qa/QA__scoring_formula_tests.sql
+++ b/db/qa/QA__scoring_formula_tests.sql
@@ -643,3 +643,4 @@ FROM (
 WHERE
     NOT ((api_score_explanation(p.product_id))->'summary') ? 'conflicts'
     OR NOT ((api_score_explanation(p.product_id))->'summary') ? 'qualified_headline';
+
diff --git a/docs/PRODUCTION_DATA.md b/docs/PRODUCTION_DATA.md
index 8c6281a3..6610173f 100644
--- a/docs/PRODUCTION_DATA.md
+++ b/docs/PRODUCTION_DATA.md
@@ -45,7 +45,7 @@
 
 ### 1.2 Migration Inventory
 
-**Location:** `supabase/migrations/` — **231 migration files**, append-only.
+**Location:** `supabase/migrations/` — **232 migration files**, append-only.
 
 **Naming convention:** `YYYYMMDDHHMMSS_description.sql` (Supabase CLI timestamps). Files are applied in lexicographic sort order.
 
@@ -329,7 +329,7 @@ There is no standalone `init_db_structure.py` script. Database initialization fo
 
 ```
 supabase db reset
-  → Applies all 231 migrations in order (supabase/migrations/*.sql)
+  → Applies all 232 migrations in order (supabase/migrations/*.sql)
   → Runs seed.sql (empty — no-op)
   → Schema is ready
 
@@ -511,7 +511,7 @@ Backup = supabase/migrations/*.sql + db/pipelines/*.sql
 ```
 
 Since the database can be fully reconstructed from:
-1. 231 migration files (schema + functions + views)
+1. 232 migration files (schema + functions + views)
 2. 25 × 4 pipeline SQL files (all product data)
 3. `ci_post_pipeline.sql` (data fixups)
 
@@ -657,3 +657,4 @@ Supabase Auth URL configuration required:
 7. **Consider EAN-based canonical keys** to eliminate `product_id` instability across environments
 8. **Add rollback documentation** — steps to revert a bad deployment
 9. **Create staging environment** — intermediate Supabase project for pre-production validation
+
diff --git a/supabase/migrations/20260605223000_backfill_ingredient_concern_tiers.sql b/supabase/migrations/20260605223000_backfill_ingredient_concern_tiers.sql
new file mode 100644
index 00000000..06a5e0ce
--- /dev/null
+++ b/supabase/migrations/20260605223000_backfill_ingredient_concern_tiers.sql
@@ -0,0 +1,142 @@
+-- Migration: Backfill ingredient concern tiers from strict additive/E-code classification.
+-- Rollback: Manual rollback only. Restore affected ingredient_ref.concern_tier/is_additive/concern_reason and products.ingredient_concern_score from backup if needed.
+-- Backfill ingredient_ref.concern_tier using strict additive/E-code classification.
+-- Safe to re-run. Does not downgrade existing higher concern tiers.
+
+BEGIN;
+
+WITH normalized AS (
+  SELECT
+    ingredient_id,
+    name_en,
+    CASE
+      WHEN lower(name_en) ~ '^\s*e[ -]?[0-9]{3,4}[a-z]?($|[^a-z0-9])'
+      THEN regexp_replace(
+        substring(lower(name_en) from '^\s*e[ -]?[0-9]{3,4}[a-z]?'),
+        '[^a-z0-9]',
+        '',
+        'g'
+      )
+      ELSE ''
+    END AS e_code,
+    lower(name_en) AS n
+  FROM ingredient_ref
+  WHERE name_en IS NOT NULL
+),
+classified AS (
+  SELECT
+    ingredient_id,
+    CASE
+      WHEN e_code IN ('e250','e252')
+        OR n LIKE '%sodium nitrite%'
+        OR n LIKE '%potassium nitrate%'
+      THEN 3
+
+      WHEN e_code IN ('e133','e150d','e211','e220','e223','e319','e385','e407','e407a','e621','e950','e951','e955')
+        OR n LIKE '%sodium benzoate%'
+        OR n LIKE '%benzoate%'
+        OR n LIKE '%sucralose%'
+        OR n LIKE '%aspartame%'
+        OR n LIKE '%carrageenan%'
+        OR n LIKE '%sodium metabisulphite%'
+        OR n LIKE '%contains sulphites%'
+        OR n IN ('sulfite','sulfiten','sulphite','sulphites')
+      THEN 2
+
+      WHEN e_code IN (
+        'e150','e150a','e150b','e150c','e172','e200','e202',
+        'e281','e282','e338','e339','e340','e341',
+        'e420','e421','e422','e440','e450','e451','e452',
+        'e460','e466','e471','e472','e500','e501','e503',
+        'e960','e960a','e960c'
+      )
+        OR n LIKE '%potassium sorbate%'
+        OR n LIKE '%sorbate%'
+      THEN 1
+
+      ELSE 0
+    END AS target_concern_tier,
+    CASE
+      WHEN e_code IN ('e250','e252')
+        OR n LIKE '%sodium nitrite%'
+        OR n LIKE '%potassium nitrate%'
+      THEN 'Strict additive/E-code classifier: nitrite/nitrate preservative concern'
+
+      WHEN e_code IN ('e133','e150d','e211','e220','e223','e319','e385','e407','e407a','e621','e950','e951','e955')
+        OR n LIKE '%sodium benzoate%'
+        OR n LIKE '%benzoate%'
+        OR n LIKE '%sucralose%'
+        OR n LIKE '%aspartame%'
+        OR n LIKE '%carrageenan%'
+        OR n LIKE '%sodium metabisulphite%'
+        OR n LIKE '%contains sulphites%'
+        OR n IN ('sulfite','sulfiten','sulphite','sulphites')
+      THEN 'Strict additive/E-code classifier: preservative, sweetener, sulphite, or carrageenan concern'
+
+      WHEN e_code IN (
+        'e150','e150a','e150b','e150c','e172','e200','e202',
+        'e281','e282','e338','e339','e340','e341',
+        'e420','e421','e422','e440','e450','e451','e452',
+        'e460','e466','e471','e472','e500','e501','e503',
+        'e960','e960a','e960c'
+      )
+        OR n LIKE '%potassium sorbate%'
+        OR n LIKE '%sorbate%'
+      THEN 'Strict additive/E-code classifier: lower concern additive or sorbate'
+
+      ELSE NULL
+    END AS target_concern_reason
+  FROM normalized
+)
+UPDATE ingredient_ref ir
+SET
+  concern_tier = GREATEST(COALESCE(ir.concern_tier, 0), c.target_concern_tier),
+  is_additive = true,
+  concern_reason = CASE
+    WHEN ir.concern_reason IS NULL OR trim(ir.concern_reason) = ''
+    THEN c.target_concern_reason
+    ELSE ir.concern_reason
+  END
+FROM classified c
+WHERE ir.ingredient_id = c.ingredient_id
+  AND c.target_concern_tier > 0
+  AND (
+    COALESCE(ir.concern_tier, 0) < c.target_concern_tier
+    OR ir.is_additive IS DISTINCT FROM true
+    OR ir.concern_reason IS NULL
+    OR trim(ir.concern_reason) = ''
+  );
+
+WITH product_scores AS (
+  SELECT
+    pi.product_id,
+    LEAST(100, SUM(
+      CASE ir.concern_tier
+        WHEN 1 THEN 15
+        WHEN 2 THEN 40
+        WHEN 3 THEN 100
+        ELSE 0
+      END
+    ))::int AS score
+  FROM product_ingredient pi
+  JOIN ingredient_ref ir ON ir.ingredient_id = pi.ingredient_id
+  WHERE COALESCE(ir.concern_tier, 0) > 0
+  GROUP BY pi.product_id
+),
+computed_scores AS (
+  SELECT
+    p.product_id,
+    COALESCE(ps.score, 0)::int AS score
+  FROM products p
+  LEFT JOIN product_scores ps ON ps.product_id = p.product_id
+  WHERE p.is_deprecated IS NOT TRUE
+)
+UPDATE products p
+SET ingredient_concern_score = cs.score
+FROM computed_scores cs
+WHERE p.product_id = cs.product_id
+  AND p.is_deprecated IS NOT TRUE
+  AND p.ingredient_concern_score IS DISTINCT FROM cs.score;
+
+COMMIT;
+