diff --git a/api/api/views/media_views.py b/api/api/views/media_views.py
index d003feec9fd..906e5974700 100644
--- a/api/api/views/media_views.py
+++ b/api/api/views/media_views.py
@@ -105,6 +105,18 @@ def _get_request_serializer(self, request):
         return req_serializer
 
     def get_db_results(self, results):
+        """
+        Map ES hits to ORM model instances.
+
+        ORM instances have all necessary info needed for serializers whereas ES
+        hits only contain the subset of fields needed for indexing and search.
+        This function issues one query to the DB, using the ``identifier`` field
+        which is both unique and indexed, so it's quite performant.
+
+        :param results: the list of ES hits
+        :return: the corresponding list of ORM model instances
+        """
+
         identifiers = []
         hits = []
         for hit in results:
@@ -267,6 +279,10 @@ def related(self, request, identifier=None, *_, **__):
 
         serializer_context = self.get_serializer_context()
 
+        serializer_class = self.get_serializer()
+        if serializer_class.needs_db:
+            results = self.get_db_results(results)
+
         serializer = self.get_serializer(results, many=True, context=serializer_context)
         return self.get_paginated_response(serializer.data)
 
diff --git a/api/test/media_integration.py b/api/test/media_integration.py
index f77ae5af0eb..0a4563344ae 100644
--- a/api/test/media_integration.py
+++ b/api/test/media_integration.py
@@ -227,6 +227,10 @@ def get_terms_set(res):
             or result["creator"] == item["creator"]
         ), f"{terms_set} {get_terms_set(result)}/{result['creator']}-{item['creator']}"
 
+        assert result["license_version"] is not None
+        assert result["attribution"] is not None
+        assert result["creator_url"] is not None
+
 
 def sensitive_search_and_detail(media_type):
     search_res = requests.get(