diff --git a/internal/etl/etl.go b/internal/etl/etl.go
index 16e147b..9144848 100644
--- a/internal/etl/etl.go
+++ b/internal/etl/etl.go
@@ -25,7 +25,7 @@ type Extract interface {
 type Transform interface {
 
 	// Transform each raw record in one or more search records depending on the given configuration
-	Transform(records []t.RawRecord, collection config.GeoSpatialCollection, substitutionsFile string, synonymsFile string) ([]t.SearchIndexRecord, error)
+	Transform(records []t.RawRecord, collection config.GeoSpatialCollection) ([]t.SearchIndexRecord, error)
 }
 
 // Load - the 'L' in ETL. Datasource agnostic interface to load data into target database.
@@ -74,7 +74,10 @@ func ImportFile(collection config.GeoSpatialCollection, searchIndex string, file
 	}
 	defer target.Close()
 
-	transformer := newTransformer()
+	transformer, err := newTransformer(substitutionsFile, synonymsFile)
+	if err != nil {
+		return err
+	}
 
 	// import records in batches depending on page size
 	offset := 0
@@ -90,7 +93,7 @@ func ImportFile(collection config.GeoSpatialCollection, searchIndex string, file
 			break // no more batches of records to extract
 		}
 		log.Printf("extracted %d source records, starting transform", sourceRecordCount)
-		targetRecords, err := transformer.Transform(sourceRecords, collection, substitutionsFile, synonymsFile)
+		targetRecords, err := transformer.Transform(sourceRecords, collection)
 		if err != nil {
 			return fmt.Errorf("failed to transform raw records to search index records: %w", err)
 		}
@@ -123,6 +126,6 @@ func newTargetToLoad(dbConn string) (Load, error) {
 	return nil, fmt.Errorf("unsupported target database connection: %s", dbConn)
 }
 
-func newTransformer() Transform {
-	return t.Transformer{}
+func newTransformer(substitutionsFile string, synonymsFile string) (Transform, error) {
+	return t.NewTransformer(substitutionsFile, synonymsFile)
 }
diff --git a/internal/etl/etl_test.go b/internal/etl/etl_test.go
index 1f0f54f..265ed32 100644
--- a/internal/etl/etl_test.go
+++ b/internal/etl/etl_test.go
@@ -83,7 +83,7 @@ func TestImportGeoPackage(t *testing.T) {
 		{
 			name:  "import everything",
 			where: "",
-			count: 67230, // 33030*2 + substitution and synonyms combinations
+			count: 72210, // 33030*2 + substitution and synonyms combinations
 		},
 		{
 			name:  "with where clause",
diff --git a/internal/etl/testdata/synonyms.csv b/internal/etl/testdata/synonyms.csv
index 36c2589..1774fbb 100644
--- a/internal/etl/testdata/synonyms.csv
+++ b/internal/etl/testdata/synonyms.csv
@@ -1,3 +1,4 @@
 eerste,1ste
 tweede,2de
-fryslân,friesland
\ No newline at end of file
+fryslân,friesland
+oud,oude
\ No newline at end of file
diff --git a/internal/etl/transform/subst_and_synonyms.go b/internal/etl/transform/subst_and_synonyms.go
index 9d44b29..f6606a2 100644
--- a/internal/etl/transform/subst_and_synonyms.go
+++ b/internal/etl/transform/subst_and_synonyms.go
@@ -40,7 +40,8 @@ func (s SubstAndSynonyms) generate(fieldValuesByName map[string]string) []map[st
 		// Create map with for each key a slice of []values
 		fieldValuesByNameWithAllValues[key] = allValues
 	}
-	return generateAllCombinations(fieldValuesByNameWithAllValues)
+	combinations := generateAllCombinations(fieldValuesByNameWithAllValues)
+	return combinations
 }
 
 // Transform a map[string][]string into a []map[string]string using the cartesian product, i.e.
@@ -82,19 +83,31 @@ func generateCombinations(keys []string, values [][]string) []map[string]string
 
 func extendValues(input []string, mapping map[string]string) []string {
 	var results []string
-	results = append(results, input...)
 
-	for j := range input {
+	for len(input) > 0 {
+		// Pop the first element from the input slice
+		current := input[0]
+		input = input[1:]
+
+		// Add the current string to the results
+		results = append(results, current)
+
+		// Generate new strings based on the mapping
 		for oldChar, newChar := range mapping {
-			if strings.Contains(input[j], oldChar) {
-				for i := 0; i < strings.Count(input[j], oldChar); i++ {
-					extendedInput := replaceNth(input[j], oldChar, newChar, i+1)
-					subCombinations := extendValues([]string{extendedInput}, mapping)
-					results = append(results, subCombinations...)
+			if strings.Contains(current, oldChar) {
+				for i := 0; i < strings.Count(current, oldChar); i++ {
+					if strings.HasPrefix(newChar, oldChar) {
+						// skip to prevent endless loop for cases such as
+						// oldChar = "foo", newChar = "foos" and input = "foosball", which would otherwise result in "foosssssssssssssssball"
+						continue
+					}
+					extendedInput := replaceNth(current, oldChar, newChar, i+1)
+					input = append(input, extendedInput)
 				}
 			}
 		}
 	}
+
 	// Possible performance improvement here by avoiding duplicates in the first place
 	return uniqueSlice(results)
 }
diff --git a/internal/etl/transform/subst_and_synonyms_test.go b/internal/etl/transform/subst_and_synonyms_test.go
index fbe244b..4cbd025 100644
--- a/internal/etl/transform/subst_and_synonyms_test.go
+++ b/internal/etl/transform/subst_and_synonyms_test.go
@@ -7,7 +7,7 @@ import (
 	"github.com/stretchr/testify/assert"
 )
 
-func Test_generateAllFieldValues(t *testing.T) {
+func Test_generate(t *testing.T) {
 	type args struct {
 		fieldValuesByName map[string]string
 		substitutionsFile string
@@ -23,6 +23,7 @@ func Test_generateAllFieldValues(t *testing.T) {
 		{"single synonym record", args{map[string]string{"component_thoroughfarename": "eerste", "component_postaldescriptor": "1234AB", "component_addressareaname": "bar"}, "../testdata/substitutions.csv", "../testdata/synonyms.csv"}, []map[string]string{{"component_thoroughfarename": "eerste", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}, {"component_thoroughfarename": "1ste", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}}, assert.NoError},
 		{"single synonym with capital", args{map[string]string{"component_thoroughfarename": "Eerste", "component_postaldescriptor": "1234AB", "component_addressareaname": "bar"}, "../testdata/substitutions.csv", "../testdata/synonyms.csv"}, []map[string]string{{"component_thoroughfarename": "eerste", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}, {"component_thoroughfarename": "1ste", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}}, assert.NoError},
 		{"two-way synonym record", args{map[string]string{"component_thoroughfarename": "eerste 2de", "component_postaldescriptor": "1234AB", "component_addressareaname": "bar"}, "../testdata/substitutions.csv", "../testdata/synonyms.csv"}, []map[string]string{{"component_thoroughfarename": "eerste 2de", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}, {"component_thoroughfarename": "1ste 2de", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}, {"component_thoroughfarename": "eerste tweede", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}, {"component_thoroughfarename": "1ste tweede", "component_postaldescriptor": "1234ab", "component_addressareaname": "bar"}}, assert.NoError},
+		{"avoid endless loop for synonyms that contain source value", args{map[string]string{"street": "oude kerkstraat"}, "../testdata/substitutions.csv", "../testdata/synonyms.csv"}, []map[string]string{{"street": "oude kerkstraat"}, {"street": "oud kerkstraat"}}, assert.NoError},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
@@ -135,7 +136,7 @@ func Test_readCsvFile(t *testing.T) {
 		wantErr assert.ErrorAssertionFunc
 	}{
 		{"Read substitutions csv", args{"../testdata/substitutions.csv"}, map[string]string{"ae": "a", "à": "a"}, assert.NoError},
-		{"Read synonyms csv", args{"../testdata/synonyms.csv"}, map[string]string{"eerste": "1ste", "fryslân": "friesland", "tweede": "2de"}, assert.NoError},
+		{"Read synonyms csv", args{"../testdata/synonyms.csv"}, map[string]string{"eerste": "1ste", "fryslân": "friesland", "oud": "oude", "tweede": "2de"}, assert.NoError},
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
diff --git a/internal/etl/transform/transform.go b/internal/etl/transform/transform.go
index ea0e7b6..a00eae9 100644
--- a/internal/etl/transform/transform.go
+++ b/internal/etl/transform/transform.go
@@ -33,14 +33,16 @@ type SearchIndexRecord struct {
 	Bbox              *pggeom.Polygon
 }
 
-type Transformer struct{}
+type Transformer struct {
+	substAndSynonyms *SubstAndSynonyms
+}
 
-func (t Transformer) Transform(records []RawRecord, collection config.GeoSpatialCollection, substitutionsFile string, synonymsFile string) ([]SearchIndexRecord, error) {
+func NewTransformer(substitutionsFile string, synonymsFile string) (*Transformer, error) {
 	substAndSynonyms, err := NewSubstAndSynonyms(substitutionsFile, synonymsFile)
-	if err != nil {
-		return nil, err
-	}
+	return &Transformer{substAndSynonyms}, err
+}
 
+func (t Transformer) Transform(records []RawRecord, collection config.GeoSpatialCollection) ([]SearchIndexRecord, error) {
 	result := make([]SearchIndexRecord, 0, len(records))
 	for _, r := range records {
 		fieldValuesByName, err := slicesToStringMap(collection.Search.Fields, r.FieldValues)
@@ -51,7 +53,7 @@ func (t Transformer) Transform(records []RawRecord, collection config.GeoSpatial
 		if err != nil {
 			return nil, err
 		}
-		allFieldValuesByName := substAndSynonyms.generate(fieldValuesByName)
+		allFieldValuesByName := t.substAndSynonyms.generate(fieldValuesByName)
 		suggestions := make([]string, 0, len(collection.Search.ETL.SuggestTemplates))
 		for i := range allFieldValuesByName {
 			for _, suggestTemplate := range collection.Search.ETL.SuggestTemplates {