diff --git a/refresh_coordinator.go b/refresh_coordinator.go
index a037df0..2bd8d88 100644
--- a/refresh_coordinator.go
+++ b/refresh_coordinator.go
@@ -233,8 +233,14 @@ func (rc *RefreshCoordinator) CoordinateRefresh(
 	}
 
 	if isNew {
-		// We created a new operation, so we need to execute it
-		go rc.executeRefreshAsync(operation, sessionID, tokenHash, refreshFunc) //nolint:gosec // long-lived background refresh intentionally uses a background context
+		// We created a new operation, so we need to execute it. Track the
+		// goroutine so Shutdown waits for in-flight refreshes to finish (they
+		// are aborted promptly via rc.ctx when stopChan closes).
+		rc.wg.Add(1)
+		go func() {
+			defer rc.wg.Done()
+			rc.executeRefreshAsync(operation, sessionID, tokenHash, refreshFunc) //nolint:gosec // long-lived background refresh intentionally uses a background context
+		}()
 	} else {
 		// Joined existing operation - this is a deduplicated request
 		atomic.AddInt64(&rc.metrics.deduplicatedRequests, 1)
diff --git a/review_r63_coord_shutdown_test.go b/review_r63_coord_shutdown_test.go
new file mode 100644
index 0000000..0ccf3c8
--- /dev/null
+++ b/review_r63_coord_shutdown_test.go
@@ -0,0 +1,57 @@
+package traefikoidc
+
+import (
+	"context"
+	"testing"
+	"time"
+)
+
+// TestRefreshCoordinatorShutdownWaitsForInflight verifies that Shutdown waits
+// for in-flight refresh operations to finish before returning. Previously
+// in-flight refresh goroutines were not tracked, so Shutdown returned
+// immediately while a refresh was still running — leaking the goroutine and
+// letting it continue using resources after the coordinator was torn down
+// (e.g. during Traefik plugin reload).
+func TestRefreshCoordinatorShutdownWaitsForInflight(t *testing.T) {
+	logger := GetSingletonNoOpLogger()
+	cfg := DefaultRefreshCoordinatorConfig()
+	cfg.RefreshTimeout = 10 * time.Second
+	rc := NewRefreshCoordinator(cfg, logger)
+
+	started := make(chan struct{})
+	release := make(chan struct{})
+
+	go func() {
+		_, _ = rc.CoordinateRefresh(context.Background(), "s1", "rt1", func() (*TokenResponse, error) {
+			close(started)
+			<-release
+			return &TokenResponse{}, nil
+		})
+	}()
+
+	select {
+	case <-started:
+	case <-time.After(2 * time.Second):
+		t.Fatal("refresh never started")
+	}
+
+	shutDone := make(chan struct{})
+	go func() { rc.Shutdown(); close(shutDone) }()
+
+	// Shutdown must NOT return while a refresh is still in flight (old
+	// behavior returned immediately, dropping the in-flight goroutine).
+	select {
+	case <-shutDone:
+		t.Fatal("Shutdown returned while a refresh was still in flight")
+	case <-time.After(300 * time.Millisecond):
+	}
+
+	close(release)
+
+	select {
+	case <-shutDone:
+	case <-time.After(2 * time.Second):
+		t.Fatal("Shutdown did not return after the in-flight refresh completed")
+	}
+	_ = logger
+}