diff --git a/siuba/sql/verbs.py b/siuba/sql/verbs.py
index 18ce57bd..8102ebe5 100644
--- a/siuba/sql/verbs.py
+++ b/siuba/sql/verbs.py
@@ -134,8 +134,7 @@ def col_expr_requires_cte(call, sel, is_mutate = False):
 
     call_vars = set(call.op_vars(attr_calls = False))
 
-    columns = lift_inner_cols(sel)
-    sel_labs = set(k for k,v in columns.items() if isinstance(v, sql.elements.Label))
+    sel_labs = get_inner_labels(sel)
 
     # I use the acronym fwg sol (frog soul) to remember sql clause eval order
     # from, where, group by, select, order by, limit
@@ -147,6 +146,11 @@ def col_expr_requires_cte(call, sel, is_mutate = False):
             or not sel_labs.isdisjoint(call_vars)
             )
 
+def get_inner_labels(sel):
+    columns = lift_inner_cols(sel)
+    sel_labs = set(k for k,v in columns.items() if isinstance(v, sql.elements.Label))
+    return sel_labs
+
 def get_missing_columns(call, columns):
     missing_cols = set(call.op_vars(attr_calls = False)) - set(columns.keys())
     return missing_cols
@@ -465,6 +469,7 @@ def _filter(__data, *args):
     conds = []
     windows = []
     for ii, arg in enumerate(args):
+
         if isinstance(arg, Call):
             new_call = __data.shape_call(arg, verb_name = "Filter", arg_name = ii)
             #var_cols = new_call.op_vars(attr_calls = False)
@@ -690,9 +695,10 @@ def _summarize(__data, **kwargs):
                 )
 
     needs_cte = [col_expr_requires_cte(call, sel) for call in new_calls.values()]
+    group_on_labels = set(__data.group_by) & get_inner_labels(sel)
 
     # create select statement ----
-    if any(needs_cte):
+    if any(needs_cte) or len(group_on_labels):
         # need a cte, due to alias cols or existing group by
         # current select stmt has group by clause, so need to make it subquery
         cte = sel.alias()
diff --git a/siuba/tests/test_verb_summarize.py b/siuba/tests/test_verb_summarize.py
index aaa80633..2e290e04 100644
--- a/siuba/tests/test_verb_summarize.py
+++ b/siuba/tests/test_verb_summarize.py
@@ -4,7 +4,7 @@
 https://github.com/tidyverse/dbplyr/blob/master/tests/testthat/test-verb-mutate.R
 """
     
-from siuba import _, mutate, select, group_by, summarize, filter
+from siuba import _, mutate, select, group_by, summarize, filter, show_query
 from siuba.dply.vector import row_number, n
 
 import pytest
@@ -35,7 +35,7 @@ def test_summarize_ungrouped(df, query, output):
 
 
 @pytest.mark.skip("TODO: should return 1 row (#63)")
-def test_ungrouped_summarize_literal(df, query, output):
+def test_ungrouped_summarize_literal(df):
     assert_equal_query(df, summarize(y = 1), data_frame(y = 1)) 
 
 
@@ -121,3 +121,31 @@ def test_summarize_removes_series_index():
             df.assign(res = df.x + df.y).drop(columns = ["x", "y"])
             )
             
+
+@backend_sql
+def test_summarize_subquery_group_vars(backend, df):
+    query = mutate(g2 = _.g.str.upper()) >> group_by(_.g2) >> summarize(low = _.x.min())
+    assert_equal_query(
+            df,
+            query,
+            data_frame(g2 = ['A', 'B'], low = [1, 3])
+            )
+
+    # check that is uses a subquery, since g2 is defined in first query
+    text = str(query(df).last_op)
+    assert text.count('FROM') == 2
+
+
+@backend_sql
+def test_summarize_subquery_op_vars(backend, df):
+    query = mutate(x2 = _.x + 1) >> group_by(_.g) >> summarize(low = _.x2.min())
+    assert_equal_query(
+            df,
+            query,
+            data_frame(g = ['a', 'b'], low = [2, 4])
+            )
+
+    # check that is uses a subquery, since x2 is defined in first query
+    text = str(query(df).last_op)
+    assert text.count('FROM') == 2
+