diff --git a/go/vt/vttablet/tabletserver/txthrottler/tx_throttler.go b/go/vt/vttablet/tabletserver/txthrottler/tx_throttler.go
index 22b2c0799fc..620656bfa6a 100644
--- a/go/vt/vttablet/tabletserver/txthrottler/tx_throttler.go
+++ b/go/vt/vttablet/tabletserver/txthrottler/tx_throttler.go
@@ -187,7 +187,7 @@ type txThrottlerStateImpl struct {
 	healthCheck      discovery.LegacyHealthCheck
 	topologyWatchers []TopologyWatcherInterface
 
-	shardMaxLag atomic.Int64
+	shardMaxLag int64
 	endChannel  chan bool
 }
 
@@ -369,7 +369,7 @@ func (ts *txThrottlerStateImpl) throttle() bool {
 	ts.throttleMu.Lock()
 	defer ts.throttleMu.Unlock()
 
-	maxLag := ts.shardMaxLag.Load()
+	maxLag := atomic.LoadInt64(&ts.shardMaxLag)
 
 	return ts.throttler.Throttle(0 /* threadId */) > 0 &&
 		maxLag > ts.config.throttlerConfig.TargetReplicationLagSec
@@ -389,7 +389,7 @@ func (ts *txThrottlerStateImpl) updateMaxShardLag() {
 					maxLag = maxLagPerTabletType
 				}
 			}
-			ts.shardMaxLag.Store(int64(maxLag))
+			atomic.StoreInt64(&ts.shardMaxLag, int64(maxLag))
 		case _ = <-ts.endChannel:
 			break
 		}
diff --git a/go/vt/vttablet/tabletserver/txthrottler/tx_throttler_test.go b/go/vt/vttablet/tabletserver/txthrottler/tx_throttler_test.go
index b23c1a2b062..8311638e017 100644
--- a/go/vt/vttablet/tabletserver/txthrottler/tx_throttler_test.go
+++ b/go/vt/vttablet/tabletserver/txthrottler/tx_throttler_test.go
@@ -22,6 +22,7 @@ package txthrottler
 //go:generate mockgen -destination mock_topology_watcher_test.go -package txthrottler vitess.io/vitess/go/vt/vttablet/tabletserver/txthrottler TopologyWatcherInterface
 
 import (
+	"sync/atomic"
 	"testing"
 	"time"
 
@@ -148,9 +149,12 @@ func TestEnabledThrottler(t *testing.T) {
 
 	throttlerImpl, ok := throttler.state.(*txThrottlerStateImpl)
 	assert.True(t, ok)
+	// Stop the go routine that keeps updating the cached  shard's max lag to preventi it from changing the value in a
+	// way that will interfere with how we manipulate that value in our tests to evaluate different cases:
+	throttlerImpl.endChannel <- true
 
 	// 1 should not throttle due to return value of underlying Throttle(), despite high lag
-	throttlerImpl.shardMaxLag.Store(20)
+	atomic.StoreInt64(&throttlerImpl.shardMaxLag, 20)
 	assert.False(t, throttler.Throttle(100, "some-workload"))
 	assert.Equal(t, int64(1), throttler.requestsTotal.Counts()["some-workload"])
 	assert.Zero(t, throttler.requestsThrottled.Counts()["some-workload"])
@@ -175,7 +179,7 @@ func TestEnabledThrottler(t *testing.T) {
 	assert.Equal(t, int64(1), throttler.requestsThrottled.Counts()["some-workload"])
 
 	// 4 should not throttle despite return value of underlying Throttle() and priority = 100, due to low lag
-	throttlerImpl.shardMaxLag.Store(1)
+	atomic.StoreInt64(&throttlerImpl.shardMaxLag, 1)
 	assert.False(t, throttler.Throttle(100, "some-workload"))
 	assert.Equal(t, int64(4), throttler.requestsTotal.Counts()["some-workload"])
 	assert.Equal(t, int64(1), throttler.requestsThrottled.Counts()["some-workload"])